Access to this full-text is provided by Wiley.

Content available from

**International Journal of Energy Research**This content is subject to copyright. Terms and conditions apply.

Research Article

Machine Learning Algorithms for Predicting Energy

Consumption in Educational Buildings

Khaoula Elhabyb ,

1

Amine Baina,

1

Mostafa Bellafkih ,

1

and Ahmed Farouk Deifalla

2

1

National Institute of Posts and Telecommunications (INPT), Rabat, Morocco

2

Future University Cairo in Egypt, Cairo, Egypt

Correspondence should be addressed to Ahmed Farouk Deifalla; ahmed.deifalla@fue.edu.eg

Received 14 December 2023; Revised 19 March 2024; Accepted 26 March 2024; Published 13 May 2024

Academic Editor: Saleh N. Al-Saadi

Copyright © 2024 Khaoula Elhabyb et al. This is an open access article distributed under the Creative Commons Attribution

License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is

properly cited.

In the past few years, there has been a notable interest in the application of machine learning methods to enhance energy eﬃciency

in the smart building industry. The paper discusses the use of machine learning in smart buildings to improve energy eﬃciency by

analyzing data on energy usage, occupancy patterns, and environmental conditions. The study focuses on implementing and

evaluating energy consumption prediction models using algorithms like long short-term memory (LSTM), random forest, and

gradient boosting regressor. Real-life case studies on educational buildings are conducted to assess the practical applicability of

these models. The data is rigorously analyzed and preprocessed, and performance metrics such as root mean square error

(RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) are used to compare the eﬀectiveness of

the algorithms. The results highlight the importance of tailoring predictive models to the speciﬁc characteristics of each

building’s energy consumption.

1. Introduction

Artiﬁcial intelligence is rapidly being integrated into various

industries, such as healthcare, ﬁnance, and smart grids.

Among these human-centric applications, the use of AI in

smart buildings has attracted signiﬁcant attention from a

large community [1]. Smart buildings, which have been a

subject of research since the 1980s, utilize advanced technol-

ogy, data analytics, and automation systems to optimize

operations, enhance occupant comfort and productivity,

and reduce costs and energy consumption [2]. These build-

ings incorporate sensors, devices, and control systems to

monitor lighting, HVAC systems, security, and access con-

trols. Real-time data on occupancy, temperature, air quality,

and energy use can be analyzed to identify optimization

opportunities. The primary aim is to create an eﬃcient, com-

fortable, and sustainable environment for residents while

reducing costs and ecological impact.

The smart building industry is experiencing signiﬁcant

growth as society becomes more connected and digital.

According to statistics from MarketsandMarkets [3], the

industry is projected to expand at a compound annual

growth rate (CAGR) of 10.5% between 2020 and 2025,

reaching a value of $108.9 billion. This growth is driven

by factors such as increased energy usage and expenses,

advancements in machine learning and the Internet of

Things (IoT), the push for net zero energy buildings, and

regulatory changes that encourage the adoption of smart

building systems and services. Figure 1 presents the fore-

casted global market side from 2020 to 2030. Expanding

on the ﬁndings of the Zion Marketing research study [4],

it reveals the market value of 40,760 million in 2016, with

projections of a substantial growth trajectory to 61,900

million by 2024, with a CAGR exceeding 34%. This indi-

cates a rapid expansion within the market, indicating

robust trends and signiﬁcant economic development dur-

ing the study period.

The AI sector being discussed is experiencing signiﬁ-

cant growth due to the integration of the Internet of

Things (IoT) and machine learning (ML). IoT sensors col-

lect data about buildings and occupants, such as tempera-

ture, humidity, occupancy, and electricity consumption.

Hindawi

International Journal of Energy Research

Volume 2024, Article ID 6812425, 19 pages

https://doi.org/10.1155/2024/6812425

This data is centralized for optimizing building operations,

improving resident comfort, and reducing energy usage.

ML, on the other hand, is a powerful tool for processing

large amounts of data from various sources. It analyzes

this data to identify patterns and predict future events,

such as equipment failures, enabling preventative mainte-

nance [5].

The American Council for Energy-Eﬃcient Economy

(ACEEE) [6] suggests that commercial buildings can signif-

icantly reduce their energy bills by up to 30% by imple-

menting energy-eﬃcient technologies such as smart

thermostats and controlled lighting. The US Department of

Energy [7] reports that commercial buildings account for a

signiﬁcant portion of total energy consumption and green-

house gas emissions in the US. This highlights the impor-

tance of buildings that can predict energy consumption

and plan eﬃciently to reduce energy usage. Intel research

[8] indicates also that energy consumption prediction has

the potential to achieve operational cost savings, staﬀpro-

ductivity gains, and energy usage reductions. Given these

ﬁndings, the primary emphasis will be on forecasting the

energy usage of smart buildings, with a speciﬁc focus on

educational facilities, which will be analyzed for the ﬁrst

time. Understanding and predicting energy consumption

in educational environments are paramount for optimizing

resource allocation, implementing eﬀective eﬃciency mea-

sures, and establishing sustainable and cost-eﬀective opera-

tional procedures [9]. By focusing on this sector, valuable

insights can be gained to inform strategies for enhancing

energy eﬃciency and sustainability in educational buildings,

ultimately contributing to improved resource management

and environmental conservation eﬀorts.

The research concentrates on energy management

within smart buildings, aiming to forecast power consump-

tion through three distinct approaches: a traditional statisti-

cal approach employing the random forest algorithm, a deep

learning approach utilizing long short-term memory

(LSTM), and a hybrid approach leveraging the gradient

boosting regressor algorithm. These three techniques were

chosen to investigate a research gap regarding to the major-

ity of data-driven methodologies. While signiﬁcant progress

has been made in this area, limited attention has been given

to utilizing streaming and temporal data for forecasting

buildings’energy demand. This gap will be addressed

through the utilization of real historical electricity data. The

used data is analyzed to evaluate model performance and

accuracy, aiming to identify the most eﬀective approach for

smart building energy management. The research is aimed

at optimizing forecasting techniques through rigorous

comparative analysis, leveraging the strengths of LSTM, RF,

and GBR models. The study highlights the importance of

advanced machine learning in shaping smart building strate-

gies and is aimed at enhancing sustainability and eﬃciency in

energy usage. Insights from this research will inform future

advancements in energy management practices for sustain-

able development. The article is structured into several delin-

eated sections, each serving a speciﬁc purpose:

(i) Introduction: This section introduces the applica-

tion of AI within the smart building sector, setting

the context for the study

(ii) Literature analysis: Here, a comparative examina-

tion of various ML algorithms used for energy pre-

diction in smart building systems is provided,

drawing insights from existing research

(iii) Methodology: This section outlines the systematic

approach adopted in the study, encompassing data

analytics, model development, and model evalua-

tion processes

(iv) Results and discussions: Findings obtained from the

methodology are presented, followed by a compara-

tive analysis that juxtaposes these results with prior

research initiatives

(v) Conclusion: This section synthesizes the results and

provides conclusions, oﬀering perspectives on the

implications of the study’sﬁndings for the ﬁeld of

smart building energy management

2. Literature Review

A recent study conducted by the International Energy

Agency [10] has revealed concerning levels of energy

CAGR 10.6%

Revenue (USD Mn/Bn)

$121.3 billion

$67.4 billion

2019 2020 2021 2022 2023 2024 2025

Ye a r

2026 2027 2028 2029 2030

Figure 1: The global smart building market size [3].

2 International Journal of Energy Research

consumption in buildings. The study found that buildings

are responsible for a signiﬁcant portion of electricity con-

sumption and overall energy consumption in urban areas.

Buildings account for 72% of total electricity consumption

and 38% of average energy consumption in urban areas.

Additionally, buildings contribute to almost 40% of total

carbon dioxide pollution in urban areas. A smart building

is a modern infrastructure that incorporates automated

control systems and uses data to improve the building’s

performance and occupants’comfort. Figure 2 presents

the smart building functionalities and its most important

axis of work.

The top technology companies are currently prioritizing

IoT (Internet of Things) and AI (artiﬁcial intelligence). The

future of building innovation is expected to focus on achiev-

ing maximum energy eﬃciency, and this challenge can be

addressed by integrating AI-powered systems like machine

learning (ML) and deep learning. ML systems continuously

improve themselves, leading to advancements in various AI

research areas [12]. ML involves algorithms that allow them

to respond to inputs from their environment and identify

nonlinear connections in complicated or uncertain systems.

ML is divided into four major categories based on the type of

learning task they manage: supervised learning, unsuper-

vised learning, semisupervised learning, and reinforcement

learning.

(i) Supervised learning is a method of developing a

machine learning model by using a labeled data

set. In this process, each data point in the set is asso-

ciated with a known intended output. The model is

trained to predict the output

(ii) Unsupervised learning: in contrast to traditional

supervised learning, developing a model on an unla-

beled data set involves working with data where the

target outputs are unknown. In this scenario, the

model is not explicitly instructed on what to search

for but instead learns

(iii) Semisupervised learning is a learning approach that

combines supervised and unsupervised learning. In

this approach, the model is trained using a data set

that is partly labeled, meaning that some of the data

points have known labels

(iv) Reinforcement learning where a model is trained to

make a series of decisions in a changing environ-

ment. The model learns through trial and error,

receiving feedback in the form of rewards or costs

Energy consumption prediction is a valuable technique

that involves forecasting the amount of energy a system or

device will use within a speciﬁc time frame. This technique

serves various purposes, such as optimizing energy usage,

predicting future energy demands, and identifying potential

ineﬃciencies in energy consumption. To predict energy

consumption, diﬀerent methods can be employed, including

statistical models, machine learning algorithms, and physics-

based models. The choice of technique depends on factors

such as data availability, system complexity, and the desired

level of accuracy. In this particular case, the focus is on

utilizing machine learning algorithms to predict energy

consumption by leveraging historical data and other relevant

factors.

The quality and relevance of the data used in machine

learning algorithms greatly inﬂuence their performance. In

a study conducted by Ahajjam et al. [13] on Moroccan

Buildings’Electricity Consumption Data Set, electricity con-

sumption was categorized into three types: whole premises

(WP), individual loads (IL), and circuit-level (CL) data.

(1) Labeled WP: Labeled whole premises (WP) con-

sumption data refers to electricity usage data col-

lected from 13 households in the MORED data set.

Energy

Smart meters,

demand response Wat er

Smart meters, use

and ow sensing

HVAC

Fans, variable air

volume, air quality

Elevators

Maintenance,

performance

Access and security

Badge in, cameras,

integration perimeter,

doors

Lighting

Occupancy sensing

Fire

Functionality checks,

detector service

24/7 monitoring

Condition monitoring,

parking lot utilization

PEHV charging

Charging of hybrid

and electric vehicles

Figure 2: The global smart building market size [11].

3International Journal of Energy Research

This data is valuable as it includes not only the raw

electricity consumption measurements but also addi-

tional information that can assist in analyzing,

modeling, and comprehending the patterns of elec-

tricity usage in diﬀerent households

(2) Labeled IL: Ground-truth electricity consumption

refers to the electricity consumption data of individ-

ual loads (IL) that have been labeled or annotated

with accurate information. This involves recording

and labeling the operational states of speciﬁc loads,

such as refrigerators or air conditioners when they

are turned on or oﬀat speciﬁc times. Having this

ground-truth information is valuable for researchers

and analysts as it allows for accurate load disaggrega-

tion, energy management, and appliance recognition

(3) CL: Measurements in the context of energy refer to

the circuit-level energy measurements obtained from

the electrical mains of a premises. These measure-

ments provide information about the overall energy

consumption of a circuit and can be used to under-

stand the energy consumption of a group of loads

The current work focuses on three educational buildings

located at Down Town University. Further information

about these buildings will be provided next. The subsequent

section presents a literature review on energy consumption

forecasting in various buildings using multiple machine

learning algorithms.

2.1. Traditional Machine Learning Approach. ML algorithms

have been utilized to tackle the primary challenges of

physics-driven methods in load prediction. For instance,

Somu et al. [14] developed eDemand, a new building energy

use forecasting model, using long short-term memory net-

works and an improved sine cosine optimization algorithm,

and as a result, the model outperformed previous state-of-

the-art models in real-time energy load prediction. Next,

Suranata et al. [15] focused on predicting energy consump-

tion in kitchens. They used a feature engineering technique

and a short-term memory (LSTM) model. Principal compo-

nent analysis (PCA) was applied to extract important fea-

tures, and the LSTM model was used on two tables. In

addition, Shapi et al. [16] developed a prediction model for

energy demand making use of the Microsoft Azure cloud-

based machine learning framework, The methodology of

the prediction model is provided using three distinct tech-

niques, including support vector machine, artiﬁcial neural

network, and k-nearest neighbors. The study focuses on

real-world applications in Malaysia, with two tenants from

an industrial structure chosen as case studies. The experi-

mental ﬁndings show that each tenant’s energy consumption

has a particular distribution pattern, and the suggested

model can accurately estimate energy consumption for each

renter. To forecast daily energy consumption based on

weather data, Faiq et al. [17] developed a new energy usage

prediction technique for institutional buildings using long

short-term memory (LSTM). The model, trained using

Malaysian Meteorological Department weather forecasting

data, outperformed support vector regression (SVR) and

Gaussian process regression (GPR) with the best RMSE

scores. The dropout method reduces overﬁtting, and Shap-

ley’s additive explanation is used for feature analysis. Accu-

rate energy consumption estimates can help detect and

diagnose system faults in buildings, aiding in energy policy

implementation. Further, Kawahara et al. [18] explore the

application of various machine learning models to predict

voltage in lithium-ion batteries. The study includes algo-

rithms such as support vector regression, Gaussian process

regression, and multilayer perceptron. The hyperparameters

of each model were optimized using 5-fold cross-validation

on training data. The data set used consists of both simula-

tion data, generated by combining driving patterns and

applying an electrochemical model, and experimental data.

The performance of the ML models was evaluated using both

simulation and experimental data, with diﬀerent data sets

created to simulate variations in state of charge distribution.

2.2. Deep Learning and Hybrid Approaches. Additionally,

various networks integrate multiple techniques to devise

data-driven approaches. These integrated mechanisms are

commonly referred to as hybrid networks. For example,

Mohammed et al. [19] focus on the application of an

intelligent control algorithm in HVAC systems to enhance

energy eﬃciency and thermal comfort. The authors pro-

pose integrating SCADA systems with an intelligent build-

ing management system to optimize heat transmission

coeﬃcients and air temperature values. Genetic algorithms

are employed to maintain user comfort while minimizing

energy consumption. Similar to [19], Aurna et al. [20]

compare the performance of ARIMA and Holt-Winters

models in predicting energy consumption data in Ohio

and Kentucky. The study ﬁnds that the Holt-Winters model

is more accurate and eﬀective for long-term forecasting. The

authors recommend further research to consider other

parameters, and environmental factors, and explore hybrid

models for better short-term load forecasting. Next, Fer-

doush et al. [21] developed a hybrid forecasting model for

time series electrical load data. The model combines random

forest and bidirectional long short-term memory methods

and was tested on a 36-month Bangladeshi electricity con-

sumption data set. The results showed that the hybrid model

outperformed standard models in terms of accuracy. The

study emphasizes the eﬀectiveness of the hybrid machine

learning approach in improving short-term load forecasting

accuracy in the dynamic electric industry. In their study,

He and Tsang [22] developed a hybrid network combining

long short-term memory (LSTM) and improved complete

ensemble empirical mode decomposition with adaptive noise

(iCEEMDAN) to optimize electricity consumption. They

divided the initial power consumption data into patterns

using iCEEMDAN and used Bayesian-optimized LSTM to

forecast each mode independently. In the same direction,

Jin et al. [23] proposed an attention-based encoder-decoder

network with Bayesian optimization for short-term electrical

load forecasting, using a gated recurrent unit recurrent neu-

ral network for time series data modeling and a temporal

attention layer for improved prediction accuracy and

4 International Journal of Energy Research

precision. Further in their study, Olu-Ajayi et al. [24] used

various machine learning techniques to predict yearly build-

ing energy consumption using a large data set of residential

buildings. The model allows designers to enter key building

design features and anticipate energy usage early in the

development process. DNN was found to be the most eﬃ-

cient predictive model, motivating building designers to

make informed choices and optimize structures. Jang et al.

[25] created three LSTM models to compare the eﬀects of

incorporating operation pattern data on prediction perfor-

mance. The model using operation pattern data performed

the best, with a CVRMSE of 17.6% and an MBE of 0.6%.

The article by Ndife et al. [26] presents a smart power

consumption forecast model for low-powered devices. The

model utilizes advanced methodologies, such as the

ConvLSTM encoder-decoder algorithm, to accurately pre-

dict power consumption trends. The performance evalua-

tion of the model demonstrates improved accuracy and

computational eﬃciency compared to traditional methods.

Also, Duong and Nam [27] developed a machine learning

system that monitors electrical appliances to improve elec-

tricity usage behavior and reduce environmental impact.

The system utilizes load and activity sensors to track energy

consumption and operating status. After three weeks of test-

ing, the system achieved a state prediction accuracy of

93.60%. In their approach, Vennila et al. [28] propose a

hybrid model that integrates machine learning and statistical

techniques to improve the accuracy of predicting solar

energy production. The model also helps in reducing place-

ment costs by emphasizing the signiﬁcance of feature selec-

tion in forecasting. In the sale context, Kapp et al. [29]

developed a supervised machine learning model to address

energy use reduction in the industrial sector. They collected

data from 45 manufacturing sites through energy audits

and used various characteristics and parameters to predict

weather dependency and production reliance. The results

showed that a linear regressor over a transformed feature

space was a better predictor than a support vector machine.

In their research, Bhol et al. [30] propose a new method for

predicting reactive power based on real power demand. They

utilize a ﬂower pollination algorithm to optimize their model

and show that it outperforms other models like GA, PSO, and

FPA. Asiri et al. [31] used an advanced deep learning model

for accurate load forecasting in smart grid systems. They use

hybrid techniques, including LSTM and CNN, feature engi-

neering, and wavelet transforms, to enhance forecasting

accuracy and eﬃciency. The results show signiﬁcant

improvements in short-term load prediction, outperforming

traditional forecasting methods.

Table 1 contains detailed information about the algo-

rithms used, performance evaluation measurements, and

the advantages and disadvantages of each approach.

3. Methodology

This research predicts power usage in three buildings of a

private research university using a data set collected from

January 2020 to January 2023. The university is known

for its excellence in education and research across various

disciplines. The buildings under study (referred to as CLAS,

NHAI, and Cronkite) are all part of the same institution

and serve distinct functions. Building CLAS, an abbrevia-

tion of Center of Law and Society, mainly consists of an

amphitheater and oﬃces, and building NHAI, which

means Nursing and Health Innovation, consists of oﬃces

and laboratories. In contrast, Cronkite consists of class-

rooms and seminar halls.

The buildings are equipped with IoT sensors connected

to power intel sockets, and the collected data is sorted on

an open-source website server [32]. The prediction method

will use three machine learning algorithms: long short-

term memory (LSTM), random forest (RF), and gradient

boosting regressor (GBR). The data will be analyzed and

prepared before being used to train and test the models.

The methodology for forecasting energy consumption

will be divided into three sections:

(1) Data analysis involves evaluating raw data to under-

stand patterns and characteristics of electrical power

consumption data.

(2) Model training trains machine learning models,

using past data to identify patterns and correlations

between input characteristics and day power use

(3) Model test models evaluation using validation met-

rics to assess their performance and accuracy.

3.1. Data Analysis

3.1.1. Data Preparation. This study focuses on the process of

data preparation in machine learning, which is time-

consuming and computationally challenging due to the pres-

ence of missing values and uneven value scales between

features. The data was prepared using two techniques: impu-

tation of missing data and standardization. The imputation

procedure was carried out using the probabilistic principal

component analysis (PPCA) approach, a maximum likeli-

hood estimate-based technique that estimates missing values

using the expectation-maximization (EM) algorithm. This

method is developed from the principal component analysis

(PCA) method, which is used for data compression or

dimensionality reduction. The resulting cleaned data was

then subjected to standardization, also known as Z-score

normalization, to ensure an even distribution of the data

above and below the mean value as shown in equation (3):

xstandardized =x−μ

σ, 1

where μrepresents the mean, σdenotes the standard devia-

tion, and xis the original data points.

3.1.2. Data Normality Analysis. This research conducted a

normality test on each renter’s data set to determine its distri-

bution. This test is crucial for model construction and is espe-

cially important for larger sample sizes. Understanding the

data set distribution can provide valuable insights into the

prediction outcome. Kurtosis measures distribution peaks,

5International Journal of Energy Research

Table 1: Previous research in ML-driven building energy use prediction.

Authors Algorithm Data set Performance evaluation Pros Cons

Somu et al.

[14]

(i) ARIMA

(ii) Genetic algorithm-LSTM

(iii) Sine cosine optimization

algorithm-LSTM

(i) The KReSIT power consumption

data set is sourced from the Indian

Institute of Technology (IIT) in

Mumbai, India.

(i) ARIMA: MAE: 0.3479;

MAPE: 21.3333; MSE: 0.1661;

RMSE: 0.4076.

(ii) Genetic algorithm-LSTM:

MAE: 0.1804; MAPE: 5.9745;

MSE: 0.0432; RMSE: 0.2073.

(iii) (ISCOA-LSTM): MAE: 0.0819;

MAPE: 4.9688; MSE: 0.0135;

RMSE: 0.1164.

(i) Improved forecasting accuracy

(ii) Improved forecasting accuracy

(iii) Real-world applicability

(i) Sensitivity to initialization

(ii) Convergence speed

Suranata

et al. [15] (i) Long short-term memory (i) NL (i) RMSE = 62 013;MAE = 26 982;

MAPE = 12 876

(i) The ability to eﬀectively predict

energy consumption patterns in

time series data.

(i) Time-consuming training

Ferdoush

et al. [21]

(i) LSTM

(ii) RF-bi-LSTM hybrid model

(iii) Bidirectional long short-term

memory (LSTM)

(i) Bangladesh Power Development

Board covered 36 months.

(i) LSTM: MSE = 0 4776;

RMSE = 0 691;MAE = 0 5578;

MAPE = 148 7.

(ii) Bi-LSTM: MSE = 0 2943;

RMSE = 0 5425;MAE = 0 4317;

MAPE = 194 80.

(iii) RF-bi-LSTM: MSE = 0 1673;

RMSE = 0 4090;MAE = 0 3070

MAPE = 193 49.

(i) Stable learning characteristics.

(ii) Moderate generalization gap in

learning loss analysis.

(i) The hybrid model may require

speciﬁc data to utilize the

strengths of random forest and

bidirectional LSTM eﬀectively.

Yaqing et al.

[22]

(i) EMD-BO-LSTM

(ii) iCEEMDAN

(i) Real power consumption data of a

university campus for 12 months

(i) EMD-BO-LSTM: MAE = 155 77;

RMSE = 203 4;MAPE = 10 41%;

R2 = 0 8478.

(ii) iCEEMDAN-BO-LSTM:

MAE = 40 841;RMSE = 59 68;

MAPE = 2 5563%;R2 = 0 986.

(i) Adaptability and eﬃciency

(ii) Enhanced prediction accuracy (i) NL

Ndife et al.

[26] (i) ConvLSTM encoder-decoder

(i) Two million measurements were

gathered over 47 months from a

residential location in Sceaux,

France.

(i) RMSE on the model: 358 kWh

RMSE on the persistence model:

465 kWh RMSE on model A:

530 kWh RMSE on model B:

450.5 kwh

(i) Improved forecast accuracy

(ii) Suitable for low-powered devices

(iii) Eﬃcient training and prediction

time

(i) Model complexity

Duong et al.

[27] (i) Multiple layer perceptron

(i) 215 data points on the power

consumption and on/oﬀstatus of

electrical devices, in Vietnam.

(i) RMSE: 10.468

(ii) MAPE: 21.563

(i) It handles large amounts of input

data well. Makes quick predictions

after training.

(i) Slow training

Faiq et al.

[17]

(i) LSTM

(ii) LSTM-RNN

(iii) CNN-LSTM

(i) Daily data from 2018 to 2021,

from the Malaysian

Meteorological Department.

(i) LSTM: RMSE = 165 20;

MAE = 572 545.

(ii) LSTM-RNN: RMSE = 263 14;

MAE = 353 38.

(iii) CNN-LSTM: RMSE = 692 14,

MSE = 1134 1.

(i) Accurate prediction of building

energy consumption

(ii) Improved energy eﬃciency

(i) Requires a signiﬁcant amount of

historical data to create an

accurate model

Bhol et al.

[29]

(i) ARIMA

(ii) Holt-Winters ﬂower

pollination algorithm

(i) Laboratory-operated critical loads

over three months.

(i) HW-GFPA: MBE = 0 42 for

validation, 0.43 for test

RMSE = 0 80

(ii) ARIMA: MBE = 0 073 for

validation, 0.016 for test

RMSE = 0 183

(i) Scalability

(ii) Optimal hyperparameter

identiﬁcation

(i) Sensitivity to kernel selection

6 International Journal of Energy Research

while skewness measures irregular probability distribution

around the mean value [33]. Equations (2) and (3) provide

formulas for skewness and kurtosis, which are essential for

understanding the data set distribution and its impact on

the prediction outcome.

Skewness = ∑N

i=1 xi−x3

N−1∗σ3,2

Kurtosis = ∑N

i=1 xi−x4

σ4,3

where nis the number of data points in the collection, xi

is the individual data points within the sample, and xis

the sample mean.

3.1.3. Feature Selection. Feature engineering is a crucial

aspect of machine learning, involving the creation of mean-

ingful data representations to enhance model performance.

It involves careful selection, transformation, and creation of

features that capture relevant information from raw data,

enhancing predictive accuracy and interoperability. Tech-

niques like principal component analysis, domain knowledge

extraction, and creative data manipulation help models

extract patterns and make accurate predictions, bridging

the gap between raw data and actionable insights.

As previously stated, our data set comprises 27 features

detailing the characteristics of the selected buildings. To

ensure optimal input for our predictive model, we employed

a feature engineering approach leveraging a tree-based

model, speciﬁcally the random forest algorithm.

3.2. Model Development. This study uses supervised machine

learning to predict energy usage using data prepared and

trained in two groups. The model employs regressive predic-

tion using random forest, LSTM, and gradient boosting

regressor. The process from data collection to model gener-

ation is depicted in Figure 3.

3.2.1. Random Forest. A random forest regressor is a

machine learning method that combines multiple decision

trees to create a predictive model for regression tasks. Each

tree is constructed using a randomly selected subset of train-

ing data and features with H x ;θk,k=1,⋯,Kwhere xrep-

resents the observed input (covariate) vector of length pwith

associated random vector X. During prediction, the regres-

sor aggregates predictions from all trees to generate the ﬁnal

output, typically the average of the individual three predic-

tion h x = 1/k∑K

k=1h x ;θk[34]. This method is com-

monly used for pattern identiﬁcation and prediction due to

its ability to learn complicated behavior, Consequently, it is

the best choice for constructing the prediction model in

the present study. In Figure 4, we present a ﬂow chart of

the random forest algorithm.

3.2.2. Long Short-Term Memory. Sepp Hochreiter and Juer-

gen Schmidhuber introduced long short-term memory

(LSTM) in 1997 as an advanced application of recurrent

neural networks. LSTM is eﬀective in processing and pre-

dicting time series data with varying durations. It captures

long-term relationships, handles variable-length sequences,

and recalls previous data, making it useful for energy con-

sumption prediction [35]. The LSTM model structure con-

sists of three layers: input, LSTM unit, and output. The

mathematical equations used in LSTM include the forget

gate, input gate, output gate, and cell state. The following

are the equations utilized in LSTM:

it=σWi·ht−1,xt+bift=σWf·ht−1,xt+bf,

Ct=ft·Ct−1+it·Ct,

ot=σWo·ht−1,xt+boht=ot· tan hC

t,

4

where xtis the input at the step t;it,ft, and otare the input,

forgot, and output vectors; gtis the candidate activation vec-

tor, and ctis the cell state at time t.

The LSTM algorithm is a powerful tool for collecting

and transmitting information across long sequences. It is

commonly used in applications such as audio recognition,

natural language processing, and time series analysis. Based

on previous research and the availability of a time series data

set, LSTM is chosen as the algorithm for predicting energy

with high precision. Figure 5 presents a ﬂowchart of LSTM.

3.2.3. Gradient Boosting Regressor. The gradient boosting

approach is an iterative method that combines weak learners

to create a strong learner by focusing on errors at every step.

It is aimed at decreasing the loss function by ﬁnding an

approximation function of the function F x that translates

xto y. This method improves prediction performance and

lowers prediction error by matching weak learner models

to the loss function [36]. The squared error function is often

used to estimate the approximation function, which is then

used to ﬁnd the ideal settings for weak learners. The gradient

boosting regressor’s mathematical equation is as follows:

yi=Fx

i+〠

M

m=1

γmhmx

i, 5

where yiis the predicted target, xiis the input features, F xi

is the ensemble model prediction, Mis the weak model, γm

is the learning rate, and hmxiis the prediction by m−th

weak model. The current research utilized gradient boosting

due to its robust predictive performance, ability to capture

complex data linkages and nonlinear patterns, and ﬂexibility

and customization capabilities. Figure 6 depicts the gradient

boost regressor algorithm’sﬂow chart.

3.3. Model Evaluation. The data set was divided into a train-

ing group (25%) and a testing group (75%). The training

group was used to train machine learning algorithms and

create predictive models for maximum consumption data.

The testing group was used to evaluate the performance of

these models. This process is illustrated in Figure 7.

The training and testing process involved a simple parti-

tioning of data to prevent overﬁtting. Machine learning algo-

rithms’predictive models were evaluated for performance

7International Journal of Energy Research

and accuracy using metrics like R2, MSE, MAE, RMSE, and

MAPE. Each measurement deﬁnition is mentioned in

Table 2.

The present research used MSE because of its sensitivity

to errors, diﬀerentiability, and simplicity of interpretation.

The use of RMSE is preferable to MSE because it yields a

more easily understandable outcome in the original units

of the dependent variable, facilitating straightforward com-

parison across data sets or models. The mean absolute error

(MAE) is a suitable metric where the quantity of errors is

more signiﬁcant than the speciﬁc direction of the mistakes,

oﬀering a clear and direct evaluation of the model’s perfor-

mance, and MAPE is particularly valuable for comparing a

model’s prediction accuracy to the scale of the actual values.

Data

preparation

Training data

(70%) Random forest

Long short term

memory

Gradient boosting

regressor

Learning

algorithm

Predictive

model

Figure 3: Process of generating predictive model after data preparation.

Start

Input data

Random subset sampling

Feature subsampling

Construct trees

Prediction by trees

Aggregation

Final prediction

End

Figure 4: Random forest algorithm ﬂowchart.

Start

Input (xt)

Memory cell

Forget gate (ft) Input gate (it)

Output gate (ot)

Hidden state (ht)

End

Candidate cell (gt)

Figure 5: Long short-term memory algorithm ﬂowchart.

No

Yes

Calculate pseudo-residuals

Fit weak learner

Start

Input dataset

Initialize ensemble

For each iteration

Final prediction

End

Aggregated predictions

Update ensemble

Figure 6: Gradient boosting regressor ﬂow chart.

8 International Journal of Energy Research

4. Results and Discussion

The experiment results were reviewed in sections, discuss-

ing the initial processing and imputation of missing data,

energy consumption prediction for each building, and per-

formance comparisons for random forest, long short-term

memory, and gradient boosting regressor models. The pre-

sentation of results follows a hierarchy, starting with the

normality test, then data preprocessing, and ﬁnally model

evaluation.

4.1. Normality Testing of Data. The evaluation is aimed at

examining the impact of data shape on predictive model

development performance, using measures of skewness and

kurtosis. Results were compiled in Table 3 to evaluate the

data’s shape and potential deviations from normal distribu-

tion. To evaluate the normality of the energy demand data,

the two values were computed using the aggregated data

from each building spanning from January 2020 to January

2023. Figure 8 also depicts the format of the data set for a

graphical examination of normality.

Based on Table 3, the data sets for the CLAS, NHAI, and

Cronkite buildings were approximately symmetrical and

skewed with bidirectional shape distribution. However, there

were some diﬀerences in the skewness values for each build-

ing. The CLAS building showed normal asymmetry due to

power consumption and KWS, with a slightly negative skew-

ness indicating a longer left tail. The CHWTON distribution

was skewed, witha skewnessof 427578, indicating a longer left

tail. The nursing and health innovation building had a pro-

nounced asymmetry, with power consumption having a posi-

tive skewness and KWS and CHWTON having a negative

skewness, indicating balanced tails. The Cronkite building

had positive skewness values, indicating a moderate right-

skewed distribution. Overall, all three data sets were approxi-

mately symmetric, skewed, and bimodal in their form density.

The kurtosis values of all three buildings in Table 3

were less than 0, indicating that their distributions were

Data

preparation

Generate

predictive values Comparison with

the actual

recorded

maximum

consumption

Training data

(30%) Root mean square

error

Mean absolute

error

Mean squared

error

Mean absolute

percentage error

Figure 7: Testing procedure for the trained predictive model.

Table 2: Performance metrics.

Algorithms Description Math form

R-squared [37]

The coeﬃcient of determination is used to determine

how much of the variance in the dependent variable can

be explained by the independent variables.

R2=1−SSres

SStot

Mean squared error [38] A regression metric used to calculate the average squared

diﬀerence between predicted and actual values. MSE = 1

n〠

n

i=1

yi−yi

2

Root mean squared error [39]

RMSE is a widely used measure for estimating the

average variance between predicted and real values in

regression tasks.

RMSE = 1

n〠

n

i=1

yi−yi

2

Mean absolute error [40]

A regression statistic used to calculate the average

absolute diﬀerence between predicted and actual

values, ignoring the direction of mistakes.

MAE = 1

n〠

n

i=1

yi−yi

Mean absolute percentage error [39]

A commonly used method for determining forecasting

error, as it measures the average absolute percent inaccuracy

for each time period less actual values divided by actual

values, making understanding it simpler due to its scaled units.

MAPE = 1

n〠

n

i=1

yi−yi

yi

× 100

9International Journal of Energy Research

platykurtic. This was also evident in Figure 8, where the

probability distribution plot had a higher tail and a larger

peak center. However, the Cronkite building had a kurto-

sis value greater than 0, indicating a leptokurtic distribu-

tion with higher variance. CLAS and NHAI had roughly

normal distributions, but CLAS had a lower mean than

the median. Department CLAS also had an almost normal

distribution but with higher skewness and kurtosis. The

CHWTON data set had a higher variation compared to

the other data sets.

Table 3: Measurements of skewness and kurtosis for the buildings.

Application area Building name Power consumption KWS CHWTON

Skewness

CLAS -0.10206 -1 0.42329

NHAI 0.017914 -1 -1

Cronkite 0.805914 -1 0.494

Kurtosis

CLAS -0.333548 -2.0 -1.019492

NHAI -0.777519 -2.0 -2

CRONKIT 2.620576 2.056333 -0.687182

0.0000

4000

0.0001

0.0002

0.0003

0.0004

Density

0.0005

0.0006

0.0007

0.0008

5000 6000 7000 8000

0.00000

–1000

0.00005

0.00010

0.00015

Density

0.00020

0.00025

0.00030

0 1000 2000 3000 4000 5000 6000 7000

0.0000

2000

0.0001

0.0002

0.0003

Density

0.0004

1000 3000

Power_consumption (KW) CHWTON

4000 5000 6000 –0.4

Density

30

25

20

15

10

5

0

–0.2 0.0 0.2 0.4

0.0000

0.0004

0.0002

0.0006

0.0008

Density

0.0010

Power_consumption (KW) CHWTON

4000 5000 6000 7000 8000 0

Density

0.00035

0.00030

0.00025

0.00020

0.00015

0.00010

0.00005

0.00000

2000 4000 6000 8000

Figure 8: Probability density for buildings CLAS, NHAI, and Cronkite.

10 International Journal of Energy Research

4.2. Data Preprocessing. Based on Figure 9, the original data

set had various scale ranges for power consumption factors

like KWS, CHWTON, voltage, and building occupants. To

verify the prediction capacity of 29 features, multiple

approaches like correlation analysis, ensemble analysis, and

tree-based models were used. After testing the mentioned

methods, the most ideal qualities for projecting energy

demand and consumption are as follows:

(1) Previous consumption patterns

(2) Calendar: weekday, month, and season

(3) Demography: A building’s population might inﬂu-

ence consumption patterns

(4) Geographical factors such as climate. People will use

more electrical appliances at hot and low tempera-

tures, respectively

The study on missing data utilized the missingness

matrix to quantify the extent of missing data and identify

rows that contained missing values. Upon analyzing

Figure 10, it is noteworthy that none of the three data sets

exhibited any missing data.

<bound method DataFrame.info of campus bldgno

0 Downtown 309 Beus center for law and society 2020 1

2020 1

2020 1

2020 1

2020 1

2022 12

2022 12

2022 12

2022 12

2023 1

. . .. . . . . .

1 Downtown 309 Beus center for law and society

2 Downtown 309 Beus center for law and society

3 Downtown 309 Beus center for law and society

4 Downtown 309 Beus center for law and society

1092 Downtown 309 Beus center for law and society

1093 Downtown 309 Beus center for law and society

1094 Downtown 309 Beus center for law and society

1095 Downtown 309 Beus center for law and society

1096 Downtown 309 Beus center for law and society

. . . . . . . . . . . .

Day Hour KW KWS CHWTONgalsgas HTmmBTU#Houses \

\

HTmmBTUlightbulbs HTmmBTUgalsgas Total#Houses Totallightbulbs

0

1

2

3

4

0

1

2

3

4

0

1

2

3

4

8905

9162

9651

10861

10305

2.550

2.702

2.710

2.452

2.467

4

5

6

7

1

2020-01-01T00:00:00.000

2020-01-02T00:00:00.000

2020-01-03T00:00:00.000

2020-01-04T00:00:00.000

2020-01-05T00:00:00.000

2020-12-28T00:00:00.000

2020-12-29T00:00:00.000

2020-12-30T00:00:00.000

2020-12-31T00:00:00.000

2020-12-32T00:00:00.000

1092

1093

1094

1095

1096

1092

1093

1094

1095

1096

11252

10343

12972

13775

10026

2.669

2.567

2.558

2.346

2.508

4

5

6

7

1

112759

124047

118887

126654

131700

43

47

45

48

50

3232

3167

3446

3504

3071

807960

791672

861333

875937

767612

135532

129250

132119

127761

119682

52

49

51

49

46

3129

3283

3360

3380

3282

782052

820649

839804

844761

820307

1

2

3

4

5

5364.07

5902.25

5915.77

5496.93

5512.42

–0.01

–0.01

–0.01

–0.01

–0.01

101

103

109

124

117

542

517

529

511

479

. . .

. . .

. . .

. . .

. . .

. . .

1092 28 4998.70 –0.01 129 451. . .

1093 29 5012.78 –0.01 118 496. . .

1094 30 4900.29 –0.01 150 476. . .

1095 31 4631.51 –0.01 160 507. . .

1096 1 4694.85 –0.01 115 527. . .

. . . . . . . . . . . . . . . . . . . . .. . .

. . .

. . . . . . . . . . . . . . .

. . . . . .. . . . . .

Totalgalsgas GHG DOW tstamp2

Figure 9: Summary of transform data set for CLAS building.

11International Journal of Energy Research

4.3. Feature Selection. Selecting the most crucial features

plays a vital role in enhancing the eﬀectiveness, stability,

and scalability of our prediction model. Through the utiliza-

tion of a feature importance assessment method, as summa-

rized in Table 4, we identiﬁed the top ﬁve inﬂuential

features: KW, KWS, CHWTON, total houses, and

CHWTONgaslas. The ranking of these features is illustrated

in Figure 11, which shows the order of their importance.

Although the initial analysis considered all 29 parameters,

the ﬁgure only highlights features that signiﬁcantly contrib-

ute to precision, ensuring a streamlined and informative

depiction.

The study is aimed at predicting energy consumption in

three educational buildings by identifying key parameters.

Through feature selection, we have identiﬁed key parameters

that signiﬁcantly impact energy usage. These include

“CHwton”or chilled water tons which measures the cooling

capacity of chilled water systems, representing the heat

energy required to melt one ton of ice in 24 hours. Addition-

ally, “KW”denotes the power consumption of electrical

equipment and lighting systems within the buildings. “Total-

lightbulb”denotes the aggregate number of light bulbs or

lamps within the buildings, crucial for various assessments.

Furthermore, aspects of HVAC systems, like “CHWTON-

galsgas,”oﬀer insights into chilled water and gas usage.

Moreover, “Combined mmBTU”measures the heat required

to raise the temperature of water by one degree Fahrenheit.

The feature selection process helps identify the most inﬂuen-

tial parameters for the predictive model, enabling more

accurate energy consumption forecasts.

4.4. Performance Evaluation and Comparison. The predic-

tion models’performance was evaluated by comparing mul-

tiple methods for each building after training and testing.

Comparative results are shown in Table 5.

Based on the performance evaluation measurements pre-

sented in Table 5, the GBR method exhibited outstanding

performance across all buildings. Notably, the determination

coeﬃcients were remarkably high, reaching 0.998 for Cron-

kite, 0.984 for CLAS, and 0.845 for NHAI. Furthermore,

the corresponding mean squared error (MSE) values were

8.148, 5.09, and 9.17, respectively. The root mean squared

error (RMSE) and mean absolute error (MAE) also sup-

ported these results, indicating that GBR outperformed

other methods and yielded the best values. Additionally,

when assessing the mean absolute percentage error (MAPE)

results, GBR surpassed the other methods, demonstrating

the lowest error percentage. The LSTM method exhibited

lower determination coeﬃcients compared to the GBR

results, with values of 0.86 for CLAS, 0.7772 for NHAI,

and 0.7609 for Cronkite. However, when comparing LSTM

to the RF method, the performance varied across buildings.

1.0 1097

Time reviewed

0.8

0.6

0.4

0.2

0.0

1097

877

658

438

219

0

1097

Power_consumption (KW)

1097

KWS

1097

CHWTON

1097

Weekday

Figure 10: Missingness graph of CLAS building.

Table 4: Feature importance.

Features Importance

CHWTON 0.303456

CHWTONgalsgas 0.23327

Total houses 0.291235

KW 0.353657

HTmmBTU 0.083478

Combined mmBTU 0.183562

HTmmBTUgalsgas 0.285535

Total light bulbs 0.229835

12 International Journal of Energy Research

Speciﬁcally, in the Cronkite building, the random forest

method outperformed LSTM with an R2 value of 0.89. Nev-

ertheless, in terms of other metrics such as MSE and RMSE,

LSTM yielded comparably smaller values than the RF

method. Moreover, there was a signiﬁcant diﬀerence in the

MAPE results, with LSTM generating fewer errors compared

to random forest. This observation suggests that, in terms of

errors, LSTM performed better and produced a lower num-

ber of errors compared to RF. According to the forecast eval-

uation, the square error method was deemed a more suitable

evaluation metric for assessing the accuracy of the predic-

tions. Following this examination, it became clear that the

gradient boosting regressor (GBR) method performed the

best across all buildings.

Considering the data presented in Table 6, it is evident

that the algorithm closest to the real testing values is the gra-

dient boosting regressor, demonstrating good precision. The

Feature ranking

Feature

0.35

0.30

0.25

0.20

0.15

0.10

0.05

0.00

KW KWS CHWTON HTmmBTU Combined mmBTU Totalgalsgas Totallightbulbs Total#Houses CHWTONgalsgas HTmmBTUgalsgas

Importance

Figure 11: Feature importance.

Table 5: Predictions for performance evaluation using trained models.

Building Method R2 MSE RMS MAE MAPE

CLAS RF 0.8506 27.245 16.530 219.73 76.947

CLAS LSTM 0.8669 11.0921 13.3036 79.0677 56.0298

CLAS GBR 0.984 8.148 9.335 71.722 40.2587

NHAI RF 0.479 56.293 20.439 47.78 8.45123

NHAI LSTM 0.8372 27.10199 19.2844 33.0788 48.11382

NHAI GBR 0.795 15.089 17.4370 32.675 52.57254

Cronkite RF 0.89318 19.821 10.8491 117.704 56.54793

Cronkite LSTM 0.76096 26.12360 7.3153 29.09945 64.8606

Cronkite GBR 0.99817 9.1734 4.04234 16.3405 36.34167

Table 6: Real and predicted average consumption for each method.

Real values GBR test LSTM test RF test

5364.07 5511.33 5646.09 5666.75

5902.25 5881.72 5811.137 5822.69

5915.77 5900.722 5871.49 5870.67

5496.93 5491.65 5499.29 5496.55

5512.42 5523.29 5535.02 5535.18

6173.30 6178.63 6366.57 6354.28

6141.73 6296.113 6345.22 6365.09

6302.20 6364.17 6371.89 6389.07

6182.52 6182.480 6234.04 6240.37

6251.62 6171.9 6166.97 6168.348

6251.62 5606.99 5530.53 5527.52

5602,05 5626.398 5637.729 5527.52

5678,72 6769.29 6437.53 5641.79

6842,53 6893.64 6884.76 6417.95

6980,2 6701.16 6606.624 6876.06

6767,83 6695.89 6520.20 6622.32

6568,83 6394.089 6358.6 6506.90

5789,52 5730.96 5596.32934 6355.81

Table 7: Cross-validation score for models.

Algorithm RF LSTM GBR

Validation score 0.83 0.92 0.95

13International Journal of Energy Research

long short-term memory (LSTM) method follows in second

place, and the random forest algorithm comes last in terms

of accuracy in predicting average consumption. In the con-

text of result validation, K-fold cross-validation is a highly

suitable technique for our case due to its inherent advan-

tages. By partitioning the data set into K subsets, each con-

taining a representative sample of the data, K-fold cross-

validation ensures thorough training and validation of the

model. This approach maximizes data utilization and mini-

mizes bias, as every data point is utilized for both training

and validation across diﬀerent folds. Furthermore, the aver-

aging of performance metrics over multiple splits provides a

robust evaluation, eﬀectively reducing the variance associ-

ated with a single train-test split. Additionally, K-fold

cross-validation facilitates better generalization by assessing

the model’s performance across diverse subsets of the data,

ensuring that it can eﬀectively handle various scenarios. Its

utility extends to hyperparameter tuning, enabling the com-

parison of diﬀerent parameter conﬁgurations across multiple

validation sets.

In our scenario, we choose 5-fold cross-validation for its

moderate data set size, balancing computational eﬃciency

and robust performance estimation. This method ensures

reliable model evaluation without excessive computational

overhead and aligns with common practices in the ﬁeld,

allowing easier comparison with existing literature and

benchmarks. Table 7 provides the outcome of the 5-fold

cross-validation.

A line graph comparison was used to better demonstrate

the diﬀerence between the actual and anticipated average

consumption levels, as depicted in Figures 12–14. In addi-

tion, Figures 15–17 show the graphical presentation of the

regression line for the three buildings. In the CLAS and

Cronkite buildings, the gradient boosting regressor (GBR)

produces a symmetric regression line, indicating that its

predicted values closely align with the actual ones. Con-

versely, for the NHAI building data set, characterized by

nonsymmetrical data, long short-term memory (LSTM)

outperforms other models due to its ability to capture tem-

poral dependencies.

However, in the case of NHAI, the performance diﬀer-

ence between LSTM and GBR is minimal, highlighting the

suitability of both algorithms for diﬀerent data characteris-

tics. GBR excels in all cases, while LSTM’s recurrent nature

makes it valuable for handling nonlinear, time-dependent

data.

7000

6500

6000

5500

5000

4500

0 50 100 150 200 250 300

1.0

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

7000

6500

6000

5500

5000

4500

0 100 200 400 600 700 800

7500

4000

7000

6500

6000

5500

5000

4500

7000

6500

6000

5500

5000

4500

7500

4000

300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

Observed

Predicted_GBR

Observed

Predicted_LSTM

Observed

Predicted_RF

Figure 12: Real and predicted average consumption for CLAS building.

14 International Journal of Energy Research

From the analysis of all the tables and ﬁgures, we con-

clude that the best performances are consistently achieved

by the gradient boosting regressor (GBR). GBR’s sequential

training approach trains weak learners sequentially, correct-

ing errors from previous iterations, and ﬁne-tuning the

model’s predictive capabilities with each step. Additionally,

gradient descent optimization minimizes prediction errors,

leading to more accurate predictions. Following GBR, long

short-term memory (LSTM) stands out as it is speciﬁcally

designed for handling sequential data, making it well-

suited for time series forecasting and similar tasks. Its ability

to understand and process temporal patterns contributes to

accurate predictions in time-dependent scenarios. Lastly,

the random forest algorithm also delivers good results, par-

ticularly when it comes to capturing complex nonlinear cor-

relations between features and the subject variable, and its

ability to model complex interactions and patterns makes

it eﬀective.

The CLAS building has a signiﬁcantly higher energy con-

sumption rate, exceeding 30 kWh, in contrast to the other

buildings. The main reason for this diﬀerence is the large sur-

face area and the simultaneous use for many educational

objectives. On the other hand, the Cronkite building has an

energy consumption rate of 26 kWh/h, while NHAI has a

consumption rate of 12 kWh per hour. Predictive modeling

approaches are necessary for eﬃcient energy allocation and

management. Within this particular instance, the gradient

boosting regressor model demonstrates its superiority in

eﬀectively predicting outcomes for both the CLAS and Cron-

kite buildings. The choice is backed by the model’s remark-

able performance metrics, as shown by its coeﬃcient of

determination (R-squared) values of 0.99 for Cronkite and

0.98 for CLAS. This model improves the accuracy of forecast-

ing by oﬀering proactive insights into the energy needs of

each building. It also helps in preventing energy loss before

it happens and promotes eﬀorts to reduce energy usage.

5. Comparison with the Previous Study

The study compared three algorithms: random forest,

LSTM, and gradient boosting regressor, revealing their per-

formance in forecasting monthly average consumption.

8000

7500

7000

6500

6000

5500

1.0

0.8

0.6

0.4

0.2

0.0

0.7

0.6

0.5

0.4

0.2

0.0

0.3

0.1

8000

7000

6000

4000

5000

8000

7500

7000

6500

6000

5500

8000

7000

6000

4000

5000

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

Observed

Predicted_LSTM

Observed

Predicted_RF

Observed

Predicted_GBR

Figure 13: Real and predicted average consumption for Cronkite building.

15International Journal of Energy Research

The development of prediction models demonstrated their

capabilities, urging further optimization. The ﬁndings also

led to a comparative analysis with previous machine learn-

ing studies. In the ﬁrst research conducted by Khaoula

et al. in 2022 [40], four machine learning algorithms were

implemented to predict energy demand for a commercial

building over two years. The algorithms used were multiple

linear regression (MLR), long short-term memory (LSTM),

simple linear regression (LR), and random forest (RF). The

results indicated that LSTM performed the best, followed

by RF, MLR, and LR, providing valuable insights into the

regression algorithms’capabilities. In the second research,

Khaoula et al. in 2023 [41] examined energy consumption

prediction in a low-energy house over four months. Unlike

the ﬁrst research, this time, the prediction considered not

only the house’s energy but also its appliances. Three

machine learning algorithms, namely, artiﬁcial neural net-

works (ANN), recurrent neural networks (RNN), and ran-

dom forest (RF), were employed for tests. Recurrent neural

networks especially LSTM once again outperformed the

other algorithms, achieving an impressive accuracy of 96%.

RF was followed with 88% accuracy. However, ANN yielded

negative predictions, indicating its unsuitability for time

series data sets. Furthermore, in their research, Khaoula

et al. [42] used three deep learning algorithms—recurrent

neural networks (RNNs), artiﬁcial neural networks (ANNs),

and autoregressive neural networks (AR-NNs)—to forecast

the total load of HVAC systems. The results showed that

the autoregressive neural network model outperformed the

other two due to its ability to capture temporal dependencies

and patterns in time series data, which is crucial for HVAC

load prediction. AR-NNs use a simpler architecture, focus-

ing on past observations to predict future values, and their

autoregressive nature allows them to eﬀectively model the

self-dependence of time series data, leading to more accurate

predictions.

Drawing insights from these three studies, signiﬁcant

ﬁndings emerge regarding the eﬃcacy of regression algo-

rithms for energy consumption prediction. Speciﬁcally, long

short-term memory (LSTM) and random forest (RF) consis-

tently emerge as top performers, especially in handling time

series data. However, our research introduces a novel aspect

by exploring the eﬀectiveness of gradient boosting regressor

(GBR), which yielded exceptional results. Notably, GBR

5000

4500

4000

3500

3000

2500

2000

5000

4500

4000

3500

3000

2500

2000

5000

4500

4000

3500

3000

2500

2000

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

1.0

0.8

0.6

0.4

0.2

0.0

5000

4500

4000

3500

3000

2500

2000

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

0 50 100 150 200 250 300 0 100 200 400 600 700 800300 500

Observed

Predicted_RF

Observed

Predicted_LSTM

Observed

Predicted_GBR

Figure 14: Real and predicted average consumption for NHAI building.

16 International Journal of Energy Research

achieved remarkable precision, boasting an impressive accu-

racy of 98.9%. Moreover, compared to other algorithms,

GBR demonstrated superior performance with fewer errors,

as evidenced by lower root mean square (RMS), mean abso-

lute error (MAE), and mean absolute percentage error

(MAPE) values. This underscores the potential of GBR as a

formidable contender in energy consumption prediction

tasks, oﬀering a promising alternative to LSTM and RF in

certain contexts.

6. Perspectives and Future Work

For future contributions, we plan to optimize the GBR

model by increasing the data used for training and predic-

tion, which may improve eﬃciency and performance on

larger data sets. We intend also to apply a novel approach

to the gradient boosting optimizer to ﬁne-tune the model’s

parameters and hyperparameters more eﬀectively. These

eﬀorts are aimed at enhancing the GBR algorithm’s perfor-

mance for accurate energy consumption forecasting and

other applications.

Another signiﬁcant contribution of our future research

lies in the utilization of transformer models for predicting

diurnal energy consumption patterns. Transformers, origi-

nally designed for natural language processing tasks, have

shown remarkable capabilities in capturing long-range

dependencies in sequential data, making them well-suited

for time series forecasting tasks as well. By applying trans-

former architectures to predict diurnal energy consumption,

we aim to leverage their ability to eﬀectively model complex

4500

4500 5000 5500 6000

Observed elec. usage (kWh)

6500 7000

5000

5500

Predicted LSTM usage (kWh):

6000

6500

7000

4500

4000 5000 6000 7000

Observed elec. usage (kWh)

5000

5500

Predicted LSTM usage (kWh):

6000

6500

7000

4500

4500 5000 5500 6000 6500 7000

Observed elec. usage (kWh)

5000

5500

Predicted GBR (kWh):

6000

6500

7000

Figure 15: Regression line between observation and predictions for CLAS building.

5500

6000

6500

Predicted RF usage (kWh):

7000

7500

8000

5500 6000 6500 7000 7500 8000

Observed elec. usage (kWh)

4000

5000

6000

7000

8000

Predicted LSTM usage (kWh):

4000 5000 6000 7000 7000

Observed elec. usage (kWh)

4000

5000

6000

7000

8000

Predicted GBR (kWh):

4000 5000 6000 7000 8000

Observed elec. usage (kWh)

Figure 16: Regression line between observation and predictions for Cronkite building.

2000

2500

3000

3500

4000

4500

5000

Predicted RF usage (kWh):

2000 3000 4000 5000

Observed elec. usage (kWh)

2000

2500

3000

3500

4000

4500

5000

Predicted LSTM usage (kWh):

2000 3000 4000 5000

Observed elec. usage (kWh)

2000

2500

3000

3500

4000

4500

8000

Predicted GBR

2000 3000 4000 5000

Observed elec. usage (kWh)

Figure 17: Regression line between observation and predictions for NHAI building.

17International Journal of Energy Research

temporal patterns and dependencies inherent in energy con-

sumption data. Our case study focuses on commercial and

institutional buildings, where accurate energy consumption

prediction is crucial for optimizing building operations,

reducing costs, and minimizing environmental impact.

7. Conclusion

Our major focus in this research is developing an energy

consumption forecasting model given the environment of

three institutional buildings that have adopted the smart

building ecosystem. From January 2020 to January 2023,

the collected energy consumption data was subjected to sta-

tistical analysis to assess its normality. The skewness and

kurtosis values showed that the data had a variety of distri-

bution characteristics.

The predictive model development process involved data

preprocessing, which included handling missing data and

identifying feature importance. For this research’s objective,

three supervised machine learning methods, namely, gradi-

ent boosting regressor (GBR), long short-term memory

(LSTM), and random forest (RF), were selected as the algo-

rithms for the predictive model. The comparison of these

strategies was based on an assessment of their production

structures and prediction abilities. The results of our model

training and testing indicated that each strategy performed

diﬀerently for each building. Remarkably, the GBR approach

continually produced the most promising outcomes,

cementing its position as the best-performing strategy across

all three buildings: CLAS, NHAI, and Cronkite. GBR’s mean

absolute percentage error (MAPE) values were 9.337, 12.338,

and 4.045 for CLAS, NHAI, and Cronkite, respectively.

Additionally, GBR achieved a lower mean absolute error

(MAE) for CLAS and Cronkite (71.04 and 53.77, respec-

tively), while RF and LSTM yielded lower MAE results for

these two buildings. Moreover, while computing average

consumption using demand data, it was shown that the gra-

dient boosting regressor (GBR) displayed greater accuracy in

anticipating demand. This performance outperformed all

other approaches in all buildings.

In terms of future study recommendations, it is sug-

gested to use more powerful computers or platforms to run

the LSTM algorithm, potentially improving its performance.

Additionally, exploring hybrid or ensemble methods may be

beneﬁcial, as they have shown higher accuracy than single

regressors. Lastly, a comparison with another smart building

could be included to distinguish and validate the obtained

results. These recommendations can further enhance the

understanding and applicability of the energy consumption

predictive model.

Data Availability

The collected data was saved in an open-source website

server [43] and could be manually downloaded from the

platform’s website in the form of a CSV ﬁle with any sort

of aggregation [43] (https://portal.emcs.cornell.edu/d/2/

dashboard-list?orgId=2).

Conflicts of Interest

The authors declare that they have no conﬂicts of interest.

Acknowledgments

The authors thank the National Center for Scientiﬁc and

Technical Research (CNRST) for supporting and funding

this research.

References

[1] H. Farzaneh, L. Malehmirchegini, A. Bejan, T. Afolabi,

A. Mulumba, and P. P. Daka, “Artiﬁcial intelligence evolution

in smart buildings for energy eﬃciency,”Applied Sciences,

vol. 11, no. 2, p. 763, 2021.

[2] E. Khaoula, B. Amine, and B. Mostafa, “Machine Learning and

the Internet Of Things for Smart Buildings: A state of the art

survey,”in 2022 2nd International Conference on Innovative

Research in Applied Science, Engineering and Technology (IRA-

SET), pp. 1–10, Meknes, Morocco, 2022.

[3] A. Widya-Hasuti, A. Mardani, D. Streimikiene, A. Sharifara,

and F. Cavallaro, “The role of process innovation between

ﬁrm-speciﬁc capabilities and sustainable innovation in SMEs:

empirical evidence from Indonesia,”Sustainability, vol. 10,

no. 7, p. 2244, 2018.

[4] MarketsandMarkets, Smart Building Market by Component

(Solution and Services), Solution (Safety and Security Manage-

ment, Energy Management, Building Infrastructure Manage-

ment, Network Management, and IWMS), Services, Building

Type, Region - Global Forecast to 2025, 2021.

[5] K. B. Anacker, Healthy Buildings: How Indoor Spaces Drive

Performance and Productivity, J. G. Allen and J. D. Macomber,

Eds., Harvard University Press, Cambridge, 2020.

[6] American Council for an Energy-Eﬃcient Economy, “Building

technologies,”n.d., https://www.aceee.org/topics/building-

technologies.

[7] US Department of Energy, “Energy eﬃcient commercial build-

ings,”n.d., https://www.energy.gov/eere/buildings/energy-

eﬃcientcommercial-buildings.

[8] Intel, “Intelligent buildings: saving energy, making occupants

happy,”2017, https://www.intel.com/content/www/us/.

[9] Z. Chen, C. Lin, X. Zhou, L. Huang, M. Sandanayake, and

P.-S. Yap, “Recent technological advancements in BIM and

LCA integration for sustainable construction: a review,”Sus-

tainability, vol. 16, no. 3, p. 1340, 2024.

[10] P. IEA, World Energy Outlook 2022, International Energy

Agency (IEA), Paris, France, 2022.

[11] “Smart building market- by application (commercial, residen-

tial, and industrial), by automation type (energy management,

infrastructure management, network and communication

management), by service (professional service and managed

services) and by region global industry perspective, compre-

hensive analysis, and forecast 2021-2028,”https://www

.zionmarketresearch.com/report/smart-buildingmarket.

[12] S. Seyedzadeh, F. P. Rahimian, I. Glesk, and M. Roper,

“Machine learning for estimation of building energy consump-

tion and performance: a review,”Visualization in Engineering,

vol. 6, pp. 1–20, 2018.

18 International Journal of Energy Research

[13] M. A. Ahajjam, D. B. Licea, C. Essayeh, M. Ghogho, and

A. Kobbane, “MORED: a Moroccan buildings’electricity con-

sumption dataset,”Energies, vol. 13, no. 24, p. 6737, 2020.

[14] N. Somu, M. R. Gauthama Raman, and K. Ramamritham, “A

hybrid model for building energy consumption forecasting

using long short term memory networks,”Applied Energy,

vol. 261, article 114131, 2020.

[15] I. W. A. Suranata, I. N. K. Wardana, N. Jawas, and I. K. A. A.

Aryanto, “Feature engineering and long short-term memory

for energy use of appliances prediction,”TELKOMNIKA (Tele-

communication Computing Electron- ics and Control), vol. 19,

no. 3, pp. 920–930, 2021.

[16] M. K. M. Shapi, N. A. Ramli, and L. J. Awalin, “Energy con-

sumption prediction by using machine learning for smart

building: case study in Malaysia,”Developments in the Built

Environment, vol. 5, article 100037, 2021.

[17] M. Faiq, K. G. Tan, C. P. Liew et al., “Prediction of energy con-

sumption in campus buildings using long short-term mem-

ory,”Alexandria Engineering Journal, vol. 67, pp. 65–76, 2023.

[18] T. Kawahara, K. Sato, and Y. Sato, “Battery voltage prediction

technology using machine learning model with high extrapola-

tion accuracy,”International Journal of Energy Research,

vol. 2023, Article ID 5513446, 17 pages, 2023.

[19] S. A. Mohammed, O. A. Awad, and A. M. Radhi, “Optimiza-

tion of energy consumption and thermal comfort for intelli-

gent building management system using genetic algorithm,”

Indonesian Journal of Electrical Engineering and Computer Sci-

ence, vol. 20, no. 3, pp. 1613–1625, 2020.

[20] N. F. Aurna, M. T. M. Rubel, T. A. Siddiqui et al., “Time series

analysis of electric energy consumption using autoregressive

integrated moving average model and Holt Winters model,”

TELKOMNIKA (Telecommunication Computing Electronics

and Control), vol. 19, no. 3, pp. 991–1000, 2021.

[21] Z. Ferdoush, B. N. Mahmud, A. Chakrabarty, and J. Uddin,

“A short-term hybrid forecasting model for time series

electrical-load data using random forest and bidirectional

long short-term memory,”International Journal of Electrical

and Computer Engineering, vol. 11, no. 1, pp. 763–771,

2021.

[22] Y. He and K. F. Tsang, “Universities power energy manage-

ment: a novel hybrid model based on iCEEMDAN and Bayes-

ian optimized LSTM,”Energy Reports, vol. 7, pp. 6473–6488,

2021.

[23] X.-B. Jin, W.-Z. Zheng, J.-L. Kong et al., “Deep-learning fore-

casting method for electric power load via attention-based

encoder decoder with Bayesian optimization,”Energies,

vol. 14, no. 6, p. 1596, 2021.

[24] R. Olu-Ajayi, H. Alaka, I. Sulaimon, F. Sunmola, and S. Ajayi,

“Building energy consumption prediction for residential

buildings using deep learning and other machine learning

techniques,”Journal of Building Engineering, vol. 45, article

103406, 2022.

[25] J. Jang, J. Han, and S.-B. Leigh, “Prediction of heating energy

consumption with operation pattern variables for non-

residential buildings using LSTM networks,”Energy and

Buildings, vol. 255, article 111647, 2022.

[26] A. N. Ndife, W. Rakwichian, P. Muneesawang, and Y. Mensin,

“Smart power consumption forecast model with optimized

weighted average ensemble,”IAES International Journal of

Artiﬁcial Intelligence, vol. 11, no. 3, p. 1004, 2022.

[27] V. H. Duong and N. H. Nguyen, “Machine learning algorithms

for electrical appliances monitoring system using open-source

systems,”IAES International Journal of Artiﬁcial Intelligence,

vol. 11, no. 1, p. 300, 2022.

[28] C. Vennila, T. Anita, T. Sri Sudha et al., “Forecasting solar

energy production using machine learning,”International

Journal of Photoenergy, vol. 2022, Article ID 7797488, 7 pages,

2022.

[29] S. Kapp, J.-K. Choi, and T. Hong, “Predicting industrial build-

ing energy consumption with statistical and machine learning

models informed by physical system parameters,”Renewable

and Sustainable Energy Reviews, vol. 172, article 113045, 2023.

[30] R. Bhol, S. C. Swain, R. Dash, K. J. Reddy, C. Dhanamjayulu,

and B. Khan, “Short-term reactive power forecasting based

on real power demand using Holt-Winters’model ensemble

by global ﬂower pollination algorithm for microgrid,”Interna-

tional Journal of Energy Research, vol. 2023, Article ID

9733723, 22 pages, 2023.

[31] M. M. Asiri, G. Aldehim, F. A. Alotaibi, M. M. Alnﬁai,

M. Assiri, and A. Mahmud, “Short-Term Load Forecasting in

Smart Grids Using Hybrid Deep Learning,”IEEE Access,

vol. 12, pp. 23504–23513, 2024.

[32] “Real-time electricity consumption,”https://portal.emcs

.cornell.edu/d/2/dashboard-list?orgId=2.

[33] P. Mishra, C. M. Pandey, U. Singh, A. Gupta, C. Sahu, and

A. Keshri, “Descriptive statistics and normality tests for statis-

tical data,”Annals of Cardiac Anaesthesia, vol. 22, no. 1,

pp. 67–72, 2019.

[34] M. R. Segal, Machine learning benchmarks and random forest

regression, 2004.

[35] Y. Yu, X. Si, C. Hu, and J. Zhang, “A review of recurrent neural

networks: LSTM cells and network architectures,”Neural

Computation, vol. 31, no. 7, pp. 1235–1270, 2019.

[36] A. Natekin and A. Knoll, “Gradient boosting machines, a tuto-

rial,”Frontiers in Neurorobotics, vol. 7, p. 21, 2013.

[37] A. Di Bucchianico, “Coeﬃcient of determination (R

2

),”Ency-

clopedia of Statistics in Quality and Reliability, 2008.

[38] D. Wallach and B. Goﬃnet, “Mean squared error of prediction

as a criterion for evaluating and comparing system models,”

Ecological Modelling, vol. 44, no. 3-4, pp. 299–306, 1989.

[39] P. Goodwin and R. Lawton, “On the asymmetry of the sym-

metric MAPE,”International Journal of Forecasting, vol. 15,

no. 4, pp. 405–408, 1999.

[40] E. Khaoula, B. Amine, and B. Mostafa, “Evaluation and com-

parison of energy consumption prediction models,”in Inter-

national Conference on Advanced Technologies for Humanity,

Marrakech, Morocco, 2022.

[41] E. Khaoula, B. Amine, and B. Mostafa, “Evaluation and com-

parison of energy consumption prediction models case study:

smart home,”in The International Conference on Artiﬁcial

Intelligence and Computer Vision, pp. 179–187, Springer

Nature Switzerland, Cham, 2023.

[42] E. Khaoula, B. Amine, and B. Mostafa, “Forecasting diurnal

Heating Energy Consumption of HVAC system using ANN,

RNN, ARNN,”in 2023 14th International Conference on Intel-

ligent Systems: Theories and Applications (SITA), pp. 1–6,

Casablanca, Morocco, 2023.

[43] “Real time utility use data,”https://portal.emcs.cornell.edu/d/

2/dashboard-list?orgId=2.

19International Journal of Energy Research

Available via license: CC BY 4.0

Content may be subject to copyright.