Vassilis Assimakopoulos

National Technical University of Athens | NTUA · School of Electrical and Computer Engineering

Although it is generally accepted that greater forecasting accuracy can contribute towards better inventory performance, this relationship may be weak, also depending on the particularities of the products being forecast, the inventory policy considered, and the underlying costs. Using the time series of the M5 competition, we empirically explore t...

We present a machine learning approach for applying (multiple) temporal aggregation in time series forecasting settings. The method utilizes a classification model that can be used to either select the most appropriate temporal aggregation level for producing forecasts or to derive weights to properly combine the forecasts generated at various leve...

Inspired by the successful use of deep learning in computer vision, in this paper we introduce ForCNN, a novel deep learning method for univariate time series forecasting that mixes convolutional and dense layers in a single neural network. Instead of using conventional, numeric representations of time series data as input to the network, the propo...

Machine learning has shown great potential in various domains, but its appearance in inventory control optimization settings remains rather limited. We propose a novel inventory cost minimization framework that exploits advanced decision-tree based models to approximate inventory performance at an item level, considering demand patterns and key rep...

The purpose of this paper is to test empirically the value currently added by Deep Learning (DL) approaches in time series forecasting by comparing the accuracy of some state-of-the-art DL methods with that of popular Machine Learning (ML) and statistical ones. The paper consists of three main parts. The first part summarizes the results of a past...

Stock control is a key aspect in retail operations. Deciding for each item when an order should be placed, how many units should be ordered, and what the target service level should be, are just some of the parameters that retailers must define on a regular basis. Although shorter review periods and larger orders typically increase product availabi...

Daily SKU demand forecasting is a challenging task as it usually involves predicting irregular series that are characterized by intermittency and erraticness. This is particularly true when forecasting at low cross-sectional levels, such as at a store or warehouse level, or dealing with slow-moving items. Yet, accurate forecasts are necessary for s...

Forecasting has always been at the forefront of decision making and planning. The uncertainty that surrounds the future is both exciting and challenging, with individuals and organisations seeking to minimise risks and maximise utilities. The large number of forecasting applications calls for a diverse set of forecasting methods to tackle real-life...

Forecasting has always been at the forefront of decision making and planning. The uncertainty that surrounds the future is both exciting and challenging, with individuals and organisations seeking to minimise risks and maximise utilities. The large number of forecasting applications calls for a diverse set of forecasting methods to tackle real-life...

In this study, we present the results of the M5 “Accuracy” competition, which was the first of two parallel challenges in the latest M competition with the aim of advancing the theory and practice of forecasting. The main objective in the M5 “Accuracy” competition was to accurately predict 42,840 time series representing the hierarchical unit sales...

This paper describes the M5 “Uncertainty” competition, the second of two parallel challenges of the latest M competition, aiming to advance the theory and practice of forecasting. The particular objective of the M5 “Uncertainty” competition was to accurately forecast the uncertainty distributions of the realized values of 42,840 time series that re...

The scientific method consists of making hypotheses or predictions and then carrying out experiments to test them once the actual results have become available, in order to learn from both successes and mistakes. This approach was followed in the M4 competition with positive results and has been repeated in the M5, with its organizers submitting th...

The main objective of the M5 competition, which focused on forecasting the hierarchical unit sales of Walmart, was to evaluate the accuracy and uncertainty of forecasting methods in the field to identify best practices and highlight their practical implications. However, can the findings of the M5 competition be generalized and exploited by retail...

The M5 competition follows the previous four M competitions, whose purpose is to learn from empirical evidence how to improve forecasting performance and advance the theory and practice of forecasting. M5 focused on a retail sales forecasting application with the objective to produce the most accurate point forecasts for 42,840 time series that rep...

Supply chain management depends heavily on uncertain point forecasts of product sales. In order to deal with such uncertainty and optimize safety stock levels, methods that can estimate the right part of the sales distribution are required. Given the limited work that has been done in the field of probabilistic product sales forecasting, we propose...

Over the last 15 years, studies on hierarchical forecasting have moved away from single-level approaches towards proposing linear combination approaches across multiple levels of the hierarchy. Such combinations offer coherent reconciled forecasts, improved forecasting performance and aligned decision-making. This paper proposes a novel hierarchica...

Achieving high levels of product availability is crucial for retail firms that aim to improve the experience of their customers, enhance their profitability, and strengthen their position in the market. However, product availability is typically subject to numerous business decisions, agreements, and restrictions that directly or indirectly affect...

Cross-learning, i.e., training models using data of multiple time series instead of the single series being forecast, has been proven to be an effective strategy for improving the forecasting accuracy of machine learning methods and especially neural networks that are “data-hungry” in nature. This was also one of the major findings of the M4 and M5...

In their daily lives, people are confronted with situations where they need to form a schema of possible future scenarios and the likelihood of them occurring, be it about climate change, economic up- or downturn, or even the potential success of a romantic date. Be these issues of mundane or universal importance, this judgmental forecasting poses...

A lot of controversy exists around the choice of the most appropriate error measure for assessing the performance of forecasting methods. While statisticians argue for the use of measures with good statistical properties, practitioners prefer measures that are easy to communicate and understand. Moreover, researchers argue that the loss-function fo...

The main objective of the M5 competition, which focused on forecasting the hierarchical unit sales of Walmart, was to evaluate the accuracy and uncertainty of forecasting methods in the field in order to identify best practices and highlight their practical implications. However, whether the findings of the M5 competition can be generalized and exp...

The M5 competition follows the previous four M competitions, whose purpose is to learn from empirical evidence how to improve forecasting performance and advance the theory and practice of forecasting. M5 focused on a retail sales forecasting application with the objective to produce the most accurate point forecasts for 42,840 time series that rep...

My contributions to this voluminous publication can be found on pp 38-40 "The natural law of growth in competition" and on pp 169-170 "Dealing with logistic forecasts in practice"

The M4 competition identified innovative forecasting methods, advancing the theory and practice of forecasting. One of the most promising innovations of M4 was the utilization of cross-learning approaches that allow models to learn from multiple series how to accurately predict individual ones. In this paper, we investigate the potential of cross-l...

This paper describes the M5 "Uncertainty'' competition, the second of two parallel challenges of the latest M competition whose aim is to advance the theory and practice of forecasting. The particular objective of the M5 "Uncertainty'' competition was to precisely estimate the uncertainty distribution of the realized values of 42,840 time series th...

The M5 competition follows on from the four previous M competitions, organized by Spyros Makridakis, whose purpose has been to advance the theory and practice of forecasting. The M5 differs from the previous four ones in five ways. First, it uses hierarchical unit sales data, generously made available by Walmart, starting at the product-store level...

Accurate forecasts are vital for supporting the decisions of modern companies. Forecasters typically select the most appropriate statistical model for each time series. However, statistical models usually presume some data generation process while making strong assumptions about the errors. In this paper, we present a novel data-centric approach-'f...

Improving customer service is a key variable for increasing company's revenue, especially in stores where service quality is highly influenced by the skills and the availability of the employees. Although companies cannot directly improve the diligence, the efficiency, and the productivity of the employees, they can enhance their availability by op...

This paper describes the M5 Accuracy competition, the first of two parallel challenges of the latest M competition whose objective is to advance the theory and practice of forecasting. The M5 Accuracy competition focused on a retail sales forecasting application and extended the results of the previous four competitions by: (a) significantly expand...

The Theta method became popular due to its superior performance in the M3 forecasting competition. Since then, although it has been shown that Theta provides accurate forecasts for various types of data, being a solid benchmark to beat, limited research has been conducted to exploit its full potential and generalize its reach. This paper examines t...

Electricity price forecasting is a challenging task as it involves predicting series that are influenced by numerous variables, such as weather conditions, electricity consumption, and seasonal factors. Yet, accurate forecasts are necessary for supporting operational management and short- to mid-term planning of energy companies. Over the years, va...

Hierarchical forecasting methods have been widely used to support aligned decision-making by providing coherent forecasts at different aggregation levels. Traditional hierarchical forecasting approaches, such as the bottom-up and top-down methods, focus on a particular aggregation level to anchor the forecasts. During the past decades, these have b...

Model selection is a complex task widely examined in the literature due to the major gains in forecasting accuracy when performed successfully. To do so, many approaches have been proposed exploiting the available historical data in different ways. In-sample testing is the most common approach but is highly affected by the data and parameter estima...

Gamification is increasingly employed in learning environments as a way to increase student motivation and consequent learning outcomes. However, while the research on the effectiveness of gamification in the context of education has been growing, there are blind spots regarding which types of gamification may be suitable for different educational...

Achieving high accuracy in energy consumption forecasting is critical for improving energy management and planning. However, this requires the selection of appropriate forecasting models, able to capture the individual characteristics of the series to be predicted, which is a task that involves a lot of uncertainty. When hierarchies of load from di...

Many regions on earth face daily limitations in the quantity and quality of the water resources available. As a result, it is necessary to implement reliable methodologies for water consumption forecasting that will enable the better management and planning of water resources. This research analyses, for the first time, a large database containing...

Temporal hierarchies have been widely used during the past few years as they are capable to provide more accurate coherent forecasts at different planning horizons. However, they still display some limitations, being mainly subject to the forecasting methods used for generating the base forecasts and the particularities of the examined series. This...

Science is caught up in a replication crisis which has negative implications for published findings that cannot be reproduced by other researchers. However, such is not the case with the M4 Competition, which not only provided the means of effectively reproducing its submissions, but also preregistered ten predictions/hypotheses about its expected...

Accurate forecasts are vital for supporting the decisions of modern companies. In order to improve statistical forecasting performance, forecasters typically select the most appropriate model for each data. However, statistical models presume a data generation process, while making strong distributional assumptions about the errors. In this paper,...

The M4 Competition follows on from the three previous M competitions, the purpose of which was to learn from empirical evidence both how to improve the forecasting accuracy and how such learning could be used to advance the theory and practice of forecasting. The aim of M4 was to replicate and extend the three previous competitions by: (a) signific...

Time series forecasting typically involves selecting the most accurate forecasting model per series, a complex task which is significantly affected by data, model and parameter uncertainty. Combining forecasts of different models is an effective but computationally expensive alternative, especially when dealing with large datasets. On the other han...

Responses to discussions and commentaries (M4 forecasting competition)

This conference paper focuses on challenges and limitations related to the development and implementation of decision support systems for supply chain management. The utilization of such systems becomes mandatory as the volume of data increases and the operational constrains become more complex. Although lots of research has been conducted in the f...

Nowadays, the absolute majority of goods is transported by ship. Yet, the demand for maritime transport is continuously increasing, introducing both opportunities and risks for the shipping companies related to their business competition and operational planning. Simultaneously, data availability, algorithmic advances and computer power are exponen...

In order to evaluate the performance of new forecasting methods, forecasters typically exploit past forecasting competitions data. Through the years, numerous studies have based their conclusions on such datasets, making any mis-performing method unlikely to receive further attention. Yet, it has been reported that these datasets might not be indic...

When evaluating the performances of time series extrapolation methods, both researchers and practitioners typically focus on the average or median performance according to some specific error metric, such as the absolute error or the absolute percentage error. However, from a risk-assessment point of view, it is far more important to evaluate the d...

The M4 Competition is the continuation of three previous ones organized by Spyros Makridakis whose purpose has been to identify the most accurate forecasting method(s) for different types of predictions. M Competitions have attracted great interest in both the academic literature and among practitioners and have provided objective evidence of the m...

Over the last years, an increasing number of researchers and practitioners support the use of Artificial Intelligence (AI) and Machine Learning (ML) in several fields of application. In the area of decision making, Artificial Neural Networks (ANNs) are being used for forecasting, although limited objective evidence is available regarding their rela...

Comparing ML with statistical forecasting methods

This editorial has two parts. The first one describes a personal experience about our attempt to replicate a forecasting study, as well as the rejection of a submitted paper, in our view due to lack of objectivity. The second part discusses the need for reproducibility and replicability in forecasting research and provides suggestions for promoting...

The M4 competition is the continuation of three previous competitions started more than 45 years ago whose purpose was to learn how to improve forecasting accuracy, and how such learning can be applied to advance the theory and practice of forecasting. The purpose of M4 was to replicate the results of the previous ones and extend them into three di...

Containing tables A1 and A2, presenting the analytical results of the forecasting models used in the present study. The accuracy is evaluated per forecasting horizon first according to sMAPE, and then to MASE. (PDF)

Machine Learning (ML) methods have been proposed in the academic literature as alternatives to statistical ones for time series forecasting. Yet, scant evidence is available about their relative performance in terms of accuracy and computational requirements. The purpose of this paper is to evaluate such performance across multiple forecasting hori...

The environmental and economic impact of photovoltaic (PV) systems is continuously growing, serving as an effective alternative energy source. Yet, failures and underperformance, e.g., due to soiling and deterioration, can significantly affect PV production and shrink the capacity available. This becomes a great issue, especially when the plant is...

In this paper, we discuss how extrapolation can be advanced by using some of the most successful elements and paradigms from the forecasting literature. We propose a new hybrid method that utilises: a) the decomposition approach of the Theta method, but instead of considering a linear trend we allow for nonlinear trends, b) rather than employing th...

The published version of this working paper can be found at: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0194889

When analyzing or forecasting time series, it is common in practice to assess the effectiveness of the examined methods by testing their performance on datasets regularly studied in the literature. Yet, it has been repeatedly reported that these data collections are not representative as they include limited number of time series of specific domain...

Numerous forecasting support systems (FSSs) have been developed through the years to help companies select and implement forecasting procedures and to support managerial decisions. While the majority of these systems are off-the-shelf, the authors argue that such generic systems will not always be up to the task. Problems can arise due to lack of c...

The Theta model, well known for its performance in the M3 competition, proposes the decomposition of the seasonally adjusted series in two or more so-called Theta-lines of the same mean and slope with the original data. Given a parameter θ, the local curves of the series are modified based on the linear correlation coefficient of the data and a for...

Nowadays, forecasting is of utmost importance when it comes to planning, which is the backbone of every successful company. In order to help companies in their forecasting and decision-making procedures, many Forecasting Support Systems have been developed through the years. While these systems can be proven quite helpful in general, it seems that...

Nowadays, water companies face numerous challenges including decreasing water supply and increasing population. In order to overcome the problems arising from these challenges and optimize their services, companies must be able to predict water demand, as well as the population distribution and its consumer habits. Obtaining this kind of informatio...

Nowadays, informed decision making is conducted through innovative Information and Communication Technology (ICT) support systems. In order to utilize such ICT-based support systems fully, decision makers need suitable training. This paper proposes and evaluates the use of a Forecasting and Foresight Support System in an undergraduate course in bus...

This paper presents a methodology for predicting electrical consumption in energy-intensive commercial buildings through a range of energy performance indicators. Specifically, the most representative indicators per energy end use of the building (lighting, kitchen, refrigerators, etc.) are defined and appropriate time series forecasting methods a...