Content uploaded by Safa S. Abdul-Jabbar
Author content
All content in this area was uploaded by Safa S. Abdul-Jabbar on Oct 08, 2022
Content may be subject to copyright.
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 45
Data Analytics and Techniques: A Review
Safa S. Abdul-Jabbar1 and Alaa K. Farhan2
1Department of Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq
2Department of Computer Science, University of Technology, Baghdad, Iraq
Abstract—Big data of dierent types, such as texts and images,
are rapidly generated from the internet and other applications.
Dealing with this data using traditional methods is not practical
since it is available in various sizes, types, and processing speed
requirements. Therefore, data analytics has become an important
tool because only meaningful information is analyzed and extracted,
which makes it essential for big data applications to analyze and
extract useful information. This paper presents several innovative
methods that use data analytics techniques to improve the analysis
process and data management. Furthermore, this paper discusses
how the revolution of data analytics based on articial intelligence
algorithms might provide improvements for many applications.
In addition, critical challenges and research issues were provided
based on published paper limitations to help researchers distinguish
between various analytics techniques to develop highly consistent,
logical, and information-rich analyses based on valuable features.
Furthermore, the ndings of this paper may be used to identify the
best methods in each sector used in these publications, assist future
researchers in their studies for more systematic and comprehensive
analysis and identify areas for developing a unique or hybrid
technique for data analysis.
Index Terms—Big data analysis, Data analytics, Data
analysis, Data management, Machine learning
Every company collects a considerable amount of data from
various sources. So two prerequisites are needed to secure this
data and use techniques to extract useful information from
this data (Khoshbakht, Shiranzaei and Quadri, 2021; Farhan
and Ali, 2017). The use of big data has rapidly progressed
from a theory to a reality with the rapid progression of data
resources and the creation of companies specializing in big
data (Zheng and Guo, 2020; Do Nascimento, et al., 2021;
Mariani and Baggio, 2022). For example, clients struggle to
because the amount of data on the internet is constantly
rising. When a customer submits a query for information
or data to an Internet search engine, the result is typically
many pages. Hence, he faces the repetitious task of locating
describing this problem is called “Data Overloading” (Kan
and Klavans, 2002). Hence, the primary objective of this
decade of electronic revolution is to construct and ensure a
better manner of managing, collaborating, and developing
via the use of computer and information technology-based
knowledge and information-oriented services (Rajon, Shamim
and Arif, 2011; Russell and Norvig, 2020). The process of
analyzing and discovering hidden patterns, undiscovered
correlations, and other valuable business information from a
vast volume of data is known as big data analytics (Patel,
Singh and Kazi, 2017; Faizan, et al., 2020). Therefore, data
analytics is a crucial subject for many systems, such as those
that work with strings or information retrieval operations
(Abdul-jabbar and George, 2017). Furthermore, data analytics
can be used to check the privacy issues in social media, such
as tags and image uploading, as we can see on Flickr and
Facebook (Smith, et al., 2013; Abkenar, et al., 2020). Besides
social media applications, data analytics can provide many
video (Verma and Agrawal, 2016).
This paper has three overarching goals:
1. It will provide a brief history of data analytics techniques
and methods for documents and describe how data analytics
tools utilize the knowledge from all input documents.
2. Presents how the previous studies are based on multi-
algorithms and multi objective to optimize the traditional
methods and explain the researcher with a comprehensive
overview that helps him choose the suitable algorithms and
integrate them into a model according to the task at hand.
3. Finally, this paper also illustrates the limitations of each
proposed method to present new directions in future works.
The paper is structured as follows, Section II introduces
the proposed data analytics techniques and methods, and
present a compelling discussion and conclusions that inform
researchers on what they can learn from published research
papers mentioned in this research.
Data analysis primarily entails big data analytical
methodologies, systematic architecture, data mining, and
analysis tools. The most crucial phase in big data is data
Vol. X, No. 2 (2022), Article ID: ARO.10975. 11 pages
DOI: 10.14500/aro.10975
Received 01 May 2022; Accepted: 28 August 2022
Regular research paper: Published: 08 October 2022
Corresponding author’s e-mail: safa.s@csw.uobaghdad.edu.iq
Copyright © 2022 Safa S. Abdul-Jabbar and Alaa K. Farhan. This
is an open access article distributed under the Creative Commons
Attribution License.
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
46 http://doi.org/10.14500/aro.10975
making recommendations, and making judgments and
decision support tools that have gained popularity, such
as executive information systems and online analytical
processing. Therefore, data analysis and interpretation
complexity encourage researchers and companies to use
algorithms that process real-time data, analyze it, and
produce highly accurate analytics results. In addition, data
analysis can be used to investigate potential values where
this information can be used for business development and
performance enhancement, such as predictive analytics
that can make future predictions. Data analytics is a wide,
analysis varies depending on the type of application required
(Schwarz, Schwarz and Black, 2014; Harfouchi, et al., 2017;
Rajaraman, 2016). Hence, data analytics aims to answer three
categories of questions in general. As shown in Fig. 1, these
elucidate what happened in the past, what is happening now,
and what is anticipated (Ghavami, 2020).
As a result, processing and obtaining the necessary
information from an extensive database cost a lot of time
and processing power (Abdul Majeed, Kadhim and Subhi
Ali, 2017). Moreover, interdisciplinary investigation makes it
to conduct a large-scale reality check. Therefore, viable
research provides critical features for completing this activity
and overcoming the inaccessibility of analytical capabilities
(Kashyap, 2019).
science used to break data into individual components for
personal inspection and integrate these components to create
knowledge. Informally, Oracle and Cloudera have proposed a
seven-step “value-chain” approach for extracting value using
data analytics; these steps are as follows (Ghavami, 2020):
3. Data collecting.
4. Data cleaning.
5. Data modeling.
6. Data science team creating (i.e., building solid teams).
7. Optimize and repeat.
On the other hand, Dr. Carol Anne Hargreaves proposed
another seven steps for the business analytics process in
her data science process model, which also can be listed as
follows (Ghavami, 2020):
2. Explore the data.
3. Analyze the data.
4. Predict what is likely to happen.
6. Make a decision and measure the outcome.
7. Update the system with the results of the decision.
All kinds of data analytics processes, including the
traditional Knowledge Discovery in Databases (KDD)
process, and others such as (Mishra and Sharma, 2014), who
proposed six steps for data analytics and (Chen, Mao and Liu,
2014) suggested three primary steps only and many others.
These proposed systems depend on big data analytics tools
that provide valuable knowledge for enhancing business.
models that can be divided into the following types:
A. Advanced analytics and predictive modeling
Machine learning, data science, and predictive modeling
have grown widespread in every area where data analysis
plays a key role (Butcher and Smith, 2020). Data mining
is a sophisticated technique for evaluating large amounts
of data. There are two forms of data analytics: Supervised/
previous research studies, predictive modeling works in
results when used with unsupervised analytics than with
supervised analytics (Fan, et al., 2018). A prediction
model is built by learning a dataset with a known outcome
Martens and Provost, showed in 2013 that when predictive
models are created depending on varied and accurate
data, they can provide a performance improvement even
on a large amount of data. In this study, the researchers
trained models and made predictions on sparse datasets
test the proposed method. The proposed method can
conclude that the system with big data might be more
organizations with more data and better understanding may
Provost, 2013). The data analytics technologies can be used
in health-care systems as in 2014 when Pourhomayoun,
et al., 2014 proposed a new system for remote health
monitoring (Pourhomayoun, et al., 2014).
On the other hand, machine learning is one of the most
of knowledge-intensive automation that can be used in many
applications (Mishra and Sharma, 2014; Cearley, et al., 2018).
such as medical, roads and many other applications with
the risk of facing many problems in robustness, monitoring,
alignment and systemic safety that should be handled
(Rajpurkar, et al., 2017; Hendrycks, et al., 2021).Fig. 1. Big data analytics' temporal questions (Peter Ghavami, 2020).
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 47
As an example of using neural networks for advanced data
analytics systems, in 2017, Jain presented the implementation
details for detecting telecommunication fraud using Data
Data Mining. The proposed method depends on Microsoft
Azure’s Event Hub and Stream Analytics components for
fraud detection using a self-coded algorithm and a Data
Mining Neural Network Pattern Recognition tool. The
systems and provide a foundation for big data analytics and
mining (Jain, 2017). Furthermore, Talasila, et al. (2020)
presented a novel neural network-based method for medical
data analytics and disease prediction in 2020. They employed
and then fed them into a Recurrent Neural Network for
disease forecasting. As a result, the new technique had a
98.57% accuracy, more than the current accuracy presented
by the existing methods for the heart disease dataset (Talasila,
et al., 2020). Furthermore, dealing with big data can be
aided by deep learning, which has the potential to extract
complicated abstractions (Vu, et al., 2021). A new analytics
model for distant physiological data was proposed based on
The proposed model is decomposed into several steps. The
which collect the data and send it to the analytical system.
Then the data preprocessing and feature extraction step
should be done to the received data. Followed by data sample
model. The proposed model was evaluated using a subset
of data acquired from 600 heart failure patients through
a remote health monitoring system. The proposed model
dramatically improved prediction accuracy and performance
(Pourhomayoun, et al., 2014). Whereas in 2019, Corizzo,
Ceci and Malerba, 2019 were inspired by the goals of
several national governments (Corizzo, Ceci and Malerba,
2019). They employed recommended methodologies based
on distributed architectures, big data analytics, and predictive
modeling research domains. The results of the proposed
system give accurate predictions (temporal and geographical)
that are scalable in big data. While in 2021(Hamarashid, Saeed
and Rashid, 2021), a new paper was published to present a
novel model for predicting the next word depending on the
N-grams used to reduce the time for predicting the next word
in Kurdish dataset. The proposed model achieved results with
accuracy up to 96.3%. Also, in 2021 another research was
presented to produce a prediction model for healthcare centers
based on machine learning algorithms and analysis methods
(Moharram, Altamimi and Alshammari, 2021). In this paper,
they analyze the input data to reduce the number of training
data. Then, three machine learning algorithms were applied
the results and select the best one for the proposed system.
appointment no-shows in pediatric outpatient clinics with
year, Rocha, et al. (2021) used Principal Component Analysis
techniques and unsupervised algorithms to perform better
clustering. As a result, K-mean clustering algorithm shows the
best results for clustering operation (Rocha, et al., 2021).
B. Model accuracy and optimization
There are several optimization strategies available by
multiobjective optimization approaches (Zarchi and Attaran,
2019), (Wang, et al., 2011), (Jaouadi, et al., 2020). In 2020,
Castellanos, et al. showed how to specify, deploy and track
performance metrics in big data analytics applications
design process methodology based on the Attribute-Driven
and Architecture analysis method technique (Castellanos,
et al., 2020). Furthermore, many researchers employ the
approximation model instead of the accurate numerical
multiobjective optimization approaches in dealing with
complicated engineering issues (Choi, Cho and Kim, 2018).
Therefore, employing optimization techniques is the best
method for identifying suitable model parameters (Kumar,
et al., 2018). A data-driven predictive modeling strategy for
forecasting surface roughness in additive manufacturing is
developed to optimize the integrity of fabricated components.
Various sensors of various sorts are used to collect data on
temperature and vibration. An ensemble learning approach is
used to train the surface roughness prediction model. A subset
of these characteristics is chosen to enhance computational
complexity and accuracy rate. As a result, the proposed
model can provide accurate predicting results. At the same
time, the frequency amplitude of the build plate vibrations,
outcome (Li, et al., 2019). Whereas in the education sector,
Tran, et al. (2019) have published a paper that described the
by establishing Federated Learning over a wireless network.
communication latencies caused by learning accuracy level,
Federated Learning time, and energy consumption of mobile
user equipment. They found the globally optimal solution
solution provides exciting insights into design issues through
the ideal Federated Learning over wireless network learning
(time, accuracy, and user equipment’s energy cost) obtained
through numerical and theoretical analysis (Tran, et al.,
2019). In 2019 Zou, et al., proposed a new vehicle evaluation
prediction model (Zou, et al., 2019). This model is used to
optimize the traditional logistic regression algorithm by
studying the logistic mathematical model, designing the error
function, using the gradient descent method to discover the
enhanced, and the accuracy is maintained.
In 2020, Liu, et al. developed a new adaptive model for
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
48 http://doi.org/10.14500/aro.10975
on micro multi objective genetics to improve performance.
The optimization results further demonstrate the proposed
model’s usefulness in real-life applications. However, this
model needs more samples and local-densifying iterations to
provide reliable optimization results (Liu, et al., 2020). Also,
Intelligence model that aims to create a hybrid framework
for predicting and analyzing stress intensity factors. This
framework was built by building an adaptive neuro-fuzzy
inference system, tuned using two meta-heuristic algorithms:
genetic algorithm and particle swarm optimization. The
proposed model outperformed the other AI models for
accurate prediction, with R2 = 0.9913, RMSE = 23.6, and
MAE = 18.07. However, increasing the datasets generated
computations that include a variety of ranges and materials
might enhance prediction performance (Seghier, et al., 2020).
C. Natural language processing
Researchers in the discipline can leverage techniques
developed to appropriately and accurately analyze language.
For example, natural language techniques have computational
and deep learning techniques such as CNN are widely used in
this area (McNamara, et al., 2017; Shamsaldin, et al., 2019).
Hence, the Natural Processing techniques allow researchers
to collect and analyze data to extract the information (Rajput,
categorization is the optimization problem. This problem can
consider an analytical issue for document summarization,
prompting a group of academics to create a nature-inspired
optimization technique based on a multi-criteria optimization
increases of 31.09% (8.43%) and 18.63% (6.09%) in
ROUGE-2 (ROUGE-L) compared to the best single-objective
Gomez, Vega-Rodríguez and Pérez, 2018). In the same year
(Rashid, Mustafa and Saeed, 2018), Rashid, Mustafa and
Saeed (2018) applied a stemmer to Kurdish text documents
(KDC-4007 dataset). They used three algorithms: Support
Vector Machine, Naïve Bays, and Decision Tree, to classify
Kurdish text. After the preprocessing phase, they found
that the support vector machine achieved the best accuracy
among all the applied algorithms. In 2019, researcher
Sanchez-Gomez, Vega-Rodríguez and Pérez, 2018 continued
developing the research proposed in the previous year by
Bee Colony. The developed system was tested on several
datasets (the same datasets used in their previous research)
and evaluated the results using a variety of measures.
Consequently, the results for ROUGE-2 and ROUGE-L have
improved to between 7.37% and 40.76% and 2.59% and
11.24%, respectively (Sanchez-Gomez, et al., 2019).
On the other hand, Yadav and Chatterjee (2016) describe
the meaning of essential words in the content for text
summarization. Sentiment analysis is constantly utilized for
large-scale text data analysis and subjectivity analysis. This
study demonstrates that sentiment analysis may be used
to summarize the content, particularly for 50% (Yadav and
Chatterjee, 2016). Furthermore, researchers can employ a
lexicon-based technique to examine students’ responses.
A new algorithm has been suggested to establish teachers’
opinion results by extracting semantic meaning from students’
amount of positive or negative thoughts. This method displays
the instructors’ opinion results, categorized according to the
strength of the positive or negative sentences. However,
utilizing a lexicon approach to sentiment analysis is not
optimal because some crucial details might be lost (Aung and
Myo, 2017).
A summarization system can be designed depending on the
dataset’s similarity or dissimilarity measures. The research
summarization for text as a binary analysis issue. They
encode a potential subcategory of sentences to be included
in the summary, and then assessed using objective functions
such as the sentence’s location in the document. The results
show that good improvements were obtained depending on
the dataset used and the objective function (Saini, et al.,
2019). In 2021, another paper was proposed to perform
data analysis using state-of-the-art techniques. Using syntax
analysis, they developed a method capable of extracting
the recent Toolkit for ATM Incidence investigative process
taxonomy factors from free-text safety reports. Finally, they
modify a Data-Driven Method capable of automatically
determining the cause of the aircraft accident. The results
demonstrate that when merely elevated predictions are
considered, the model provided pilots’ contribution is around
97% accurate and 94% for ATCo (Buselli, et al., 2021). In
the same year, Vargas-Calderón, et al. (2021) presented a
model used in healthcare applications to evaluate the quality
of service in hospitals depending on client reviews. After the
text extraction and cleaning step, the model was designed
depending on multi-ML algorithms (Vargas-Calderón, et al.,
2021). Furthermore, in 2021 Hryshchenko and Yaremenko
and neural networks to categorize a batch of text data and
determine both disadvantages and advantages of each method
(Hryshchenko and Yaremenko, 2021). At the same time,
Yaremenko, Rogoza and Spitkovskyi (2021) developed a
neural network architecture that can process a large amount
of data in real-time systems and handle the determined
limitation of the applied mathematical models of the standard
Neural Networks and Naive Bays (Yaremenko, Rogoza and
Spitkovskyi, 2021).
D. Quantitative analysis (prediction and prognostics)
Quantitative analysis is concerned with quantifying and
analyzing variables to arrive at conclusions. It entails using
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 49
statistical tools to analyze numerical data for answering
questions such as who and when. Apuke published his
work on predictor measurement heterogeneity by altering
the degree of measurement error across derivation and
validation scenarios. Hence, he generated hybrid predictor
measurements using measurement error models (Apuke,
2017; Pajouheshnia, et al., 2019; Luijken, et al., 2019; and
Luijken, et al., 2020). In 2021, Admiraal, et al. used 12
quantitative features gathered from various patient situations
to train several types of machine learning algorithms. The
research results show that machine learning employing
quantitative features derived from collected data has a
better precision than visual data analysis in predicting poor
prognosis following cardiac arrest, making it a potential
alternative to visual analysis (Admiraal, et al., 2021). In
2022, Luijken, Song and Groenwold proposed a paper
to analyze the expected predictor measurement diversity
impact. In period outcome data, simulation research was
variation across validation and implementation settings. The
application of quantitative prediction error analysis was
demonstrated with an illustration of forecasting the 6-year
probability of acquiring second type diabetes with variability
in the predictor body mass-index measurement. As a result of
this paper, all situations of predictor measurement variability
resulted in the poor measurement of prediction models, and
overall accuracy was lowered. Furthermore, it increased
random predictor measurement variability (Luijken, Song
and Groenwold, 2022).
representations from large-scale data, particularly unlabeled
data, which is plentiful in Big Data (Chen and Lin, 2014),
(Najafabadi, et al., 2015). In 2020, Zhong, Yu and Ai
proposed the big data-based hierarchical deep learning
system in the context of employing deep learning for data
analytics. This system uses behavioral and content features
in the payload. When several machines are deployed, the
boost the detection rate of intrusive attacks and reduce the
E. Ensemble of models (data analytics prediction
framework)
In many real-world applications, the availability of
and eliminate duplicate and unnecessary variables from the
feature-set, particularly in high-dimensional applications.
This circumstance naturally happens in many real-life
situations when a large amount of data can be acquired
of samples is time-intensive and cannot be assumed. Many
approaches were suggested to improve accuracy in machine
learning; one of these approaches is to aggregate the output
of several learners. Ensemble Learning is a term used to
describe this approach. Bagging, boosting, stacking, and
error-correcting output are the four methods for merging
several models (Wang, et al., 2014). The learning under
individual classes. It is employed when there is a reference
value and training set with the variables to cluster (Dean,
2014). Whereas in unsupervised learning, feature selection
seeks to locate meaningful subsets of features that yield best
groupings clustering by clustering “similar” items together
using any similarity metric (Nag and Mitra, 2002), (Dy and
Brodley, 2004), (Hong, et al., 2008), (Elghazel and Aussem,
2010). In 2011, Rajon, Shamim and Arif proposed a complete
framework that designed and implemented a generic product-
independent e-market model for emerging economies. This
paper’s fundamental contribution is creating and executing a
generic e-marketplace model for emerging economies where
agriculture is widely practiced and a thriving manufacturing
sector. A comprehensive examination of the utility and
services has also been presented (Rajon, Shamim and Arif,
2011). Rajon, Shamim and Arif proposed a method based on
the random sample partition model that retains the statistical
features of the data set in each data block in 2018. They
presented the Alpha framework, which consists of three
primary layers for data administration, batch management,
analysis with Random Sample Partition blocks. The results
show that the proposed method can provide approximate
results for data analysis tasks such as data summarization and
the Alpha framework for Big Data Analysis tasks (Salloum,
et al., 2018). In the same year, Yu, et al. (2018) design a
model to demonstrate how boosting and bagging approaches
can be compared to produce better explanatory models to
prove that the ensamble approaches are more suitable for
some problems than other approaches (Yu, et al., 2018).
On the other hand, Kumar, Singh and Buyya proposed a
new ensemble learning-based workload prediction model in
2020, which makes use of excessive learning machines and
weights their estimates with a voting engine. The optimized
weights are chosen using a metaheuristic algorithm
motivated by the black hole theory. The results demonstrate
the approach’s superiority over conventional methods, with
a reduction in RMSE of up to 99.20% (Kumar, Singh and
Buyya, 2020). Whereas in 2021 a new framework based on
features modeling and ensemble learning to predict query
performance was proposed by Zaghloul, Salem and Ali-
Eldin (2021) using Machine learning algorithm attempting to
predict a performance metric based on the amount of time
elapsed and ensemble learning (Zaghloul, Salem and Ali-
Eldin, 2021).
Data analytics methods and techniques have more and
more applications in life, and performance enhancement
solutions are widely applied. It is essential to improve the
by enhancing the result accuracy and processing time; this
will be done by analyzing the input data and extracting
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
50 http://doi.org/10.14500/aro.10975
only the relevant information that the application needs.
Depending on this principle, many types of research papers
that can be used to enhance the analysis results. This paper
provides a survey of these research papers as summarized in
Table I. The research papers that proposed new methods in
the literature review section were presented in this table; to
was developed, the main research issues: Used to show
the problems that the research tries to solve, the research
techniques used to describe the method and tools that used to
to describe the most important results obtained or concluded
from the published research.
It can be seen from the summary of the research papers
presented in Table I that most researchers suggested methods
that used machine learning algorithms. Accordingly, this
paper discusses the proposed methods’ details and their
impact on the results, as shown in Table II. This table focuses
References Research scope Main research issues Main research techniques
De Fortuny, Martens, and
Provost, 2013
Big data Critical prediction jobs Multivariate Bernoulli
Naive Bayes
Organizations with more data
selection provide a better
understanding of improving the
system performance
Pourhomayoun, et al., 2014 Healthcare systems Remote health monitoring Multi-model approach Improved the prediction accuracy
and performance
Corizzo, Ceci and Malerba,
2019
Energy sector Accurate big data analytics and
predictive modeling
Multi-model approach Accurate results of predictions
with big data
Moharram, Altamimi, and
Alshammari, 2021
Healthcare systems Reduce the number of training
data
ML algorithms The best algorithm used was
Rocha, et al., 2021 Human development
indicators
Classify the departments of
peru according to their human
development index using
clustering techniques
Multi-model approach
Li, et al., 2019 Extrusion-based additive
manufacturing
Optimize the integrity of
fabricated components
ML algorithms Providing accurately predicted
results
Tran, et al., 2019 Education sector
(computation vs. communication
latencies) (Learning time vs.
energy consumption)
sub-problems
accuracy
Zou, et al., 2019 Vehicle evaluation Many iterations and training
vast amounts of data take a long
time
ML algorithms The training time is reduced, the
and the accuracy is maintained
Liu, et al., 2020 Multiobjective
optimization
High computational cost Multi-model approach The proposed model’s useful in
real-life applications
El and Ben, 2020 Stress intensity factor Prediction of stress intensity
factor
ML algorithms Provide accurate predictions
Sanchez-Gomez,
Vega-Rodríguez and
Pérez, 2018
Text summarization Find the essential information
from a document collection
Optimization algorithms Provide essential improvements
compared to the traditional
approaches
Sanchez-Gomez, et al., 2019 Text summarization Find the essential information
from a document collection
Optimization algorithms Provide improved performance
and accurate results compared
to the previous version of this
proposed model
Yadav and Chatterjee, 2016 Text summarization
summarize texts
Sentiment analysis
summarizing approach based on
Aung and Myo, 2017 Education system Analysis of students’ comment Sentiment analysis Provide good results but not
optimal because some crucial
details might be lost
Saini, et al., 2019 Text summarization Design an automatic
text-summarization
ML algorithms Provide good improvements to
the traditional approaches
Buselli, et al., 2021 Automation and digitization to
maintain safety aviation
Multi-model approach The model provides pilots’
contribution is around 97%
accurate and 94% for ATCo
Admiraal, et al., 2021 Healthcare systems electroencephalography
reactivity quantitative analysis
of neurological prognostication
following cardiac arrest
ML algorithms Provide better precision than
visual data analysis in predicting
poor prognosis following cardiac
arrest
(Contd...)
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 51
the proposed methods. It has been formulated to include: the
ML algorithms used to design the proposed method and the
techniques used to describe how these algorithms are used to
build the proposed method. Furthermore, this table illustrates
are also presented in this table to describe the challenges
and problems in the proposed methods. As we have seen
in (Admiraal, et al., 2021) and (Zhong, Yu, and Ai, 2020),
there are limitations in the number of models used to test
the proposed method, so it may have some implementation
issues when used with a large amount of data. Furthermore,
(Kumar, Singh, and Buyya, 2020), there is another issue:
people must determine the number of networks and nodes
in hidden networks. This causes the system to need human
intervention at the data entry stage and is not entirely
automated. While in (Jain, 2017), the proposed model was
designed without any dynamic implementation, which causes
the possibility of facing several problems when applied to
real-time data.
Data analytics aims to provide meaningful and relevant
information. However, many users are uncertain about
the sort of analysis to perform on data collection and
which kinds of visual data presentation are appropriate.
This paper presented a comprehensive review of data
analytics techniques to help researchers construct an
in the intentional system. This leads to utilizing these
analytics tools in the best way to provide better privacy,
rather than relying only on standard analytics tools to solve
Table I, ML algorithms have shown the highest usability
in data analysis systems than other algorithms because they
provide good accuracy with higher performance capacity
in multiple areas of industrial, commercial, agricultural,
health, education, and mining activity. As a result, these
algorithms contribute to the development, increased
employment, the contribution of mining techniques
and increased business investment. Furthermore, Fig. 2
illustrates the turnout percentage, which shows that ML
algorithms are superior to other methods. Therefore, the
ML improvements were presented in Table II to show the
Finally, it is worth mentioning that many challenges can
be faced when adopting ML algorithms in data analytics
systems. For example, processing time, computation power…
etc. However, these challenges can be addressed using
References Research scope Main research issues Main research techniques
Luijken, Song and
Groenwold, 2022
Healthcare systems
measurement heterogeneity
Quantitative prediction
error analysis
Proves that the increasing
random predictor measurement
heterogeneity will decrease the
model discrimination
Rajon, Shamim and Arif,
2011
Electronic commerce Design a general prototype for
the e-marketplace
General framework
designing tools
They are creating and developing
an e-market prototype that is
product-agnostic
Salloum, et al., 2018 Big data Making the extensive data
analysis is feasible when the
volume of data exceeds the
available computation power
ML algorithms
extensive data analysis
Kumar, Singh, and Buyya,
2020
Cloud systems These systems must allocate and
deallocate resources with low
operational cost and maintain
the quality of services
ML algorithms Provide accurate predictions by
reducing the error prediction
Zaghloul, Salem, and
Ali-Eldin, 2021
Query optimizer Attempt to predict a
performance metric
ML algorithms
performance metric
Jain, 2017 Fraud detection
for fraud detection using a
self-coded
ML algorithms
cloud analytics
Talasila, et al., 2020 Healthcare systems Provide an accurate method for
diseases prediction
ML algorithms Provide prediction accuracy up
to 98.57%
Zhong, Yu, and Ai, 2020 Intrusion detection Intrusion detection depends on
data analytics
ML algorithms Boost the detection rate of
intrusive attacks and reduce the
ML: Machine learning
Fig. 2. The percentage of turnout for each method.
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
52 http://doi.org/10.14500/aro.10975
parallel and distributed frameworks and choosing appropriate
algorithms to implement for each system.
algorithms used for designing each system. Furthermore, it
big data analytics methodologies, evaluating them in terms of
applications has been adjusted to construct how it can be used
to provide a high-quality performance. It should be noted that
this paper recognizes that the core function of machine learning
the behavior of previous data models. As a result, this paper
intends to provide simple research examining several proposed
gaps clearly in unknown information. From this examination,
we can conclude that when the ML algorithms and data
analytics self-tuning system feature selections have been used,
they will improve performance compared to other approaches
and techniques. Several paper investigations in many sectors
have demonstrated the possibility of using machine learning
algorithms in data analytics systems to improve performance
speed and accuracy.
Abdul Majeed, G., Kadhim, A. and Subhi Ali, R. (2017). Retrieving encrypted
query from encrypted database depending on symmetric encrypted cipher system
method. Diyala Journal For Pure Science, 13(1), pp.183-207.
References The used ML
algorithms
The design techniques Research limitation
Time Accuracy Processing
Moharram, Altamimi
and Alshammari, 2021
logistic regression,
tree
The optimum machine-learning method was discovered
by comparing three ML algorithms by computing recall
and prescience for each
* * * -
Li, et al., 2019 RF, AdaBoost, CART,
SVR, RR, RVFL
Data was collected depending on multiple sensors,
extracted and selected features from these data, then
applied selective algorithms to these features
* -
Zou, et al., 2019 Logistic regression Using the logistic regression algorithm based on the
gradient descent method and optimizing the sigmoid
function
* * * -
Seghier, et al., 2020 Genetic and particle
swarm algorithms
Hybrid AI model (GA and PSO) for accurate prediction
method calculations
* * * -
Saini, et al., 2019 Genetic operation
inspired form genetic
algorithm
self-organizing map (based on genetic operations)
to select a subset of sentences, then evaluate these
sentences using the objective functions
* * -
Admiraal, et al., 2021 LR, SVM, RF, NN, and
GTB algorithms
quantitative variables derived from data of 134 adult
cardiac arrest patients
* A limited number of
samples for training
operation
Salloum, et al., 2018 Hadoop MapReduce
and data mining and
machine learning
algorithms
The data is assigned to the blocks sampling algorithm,
followed by blocks analysis and ensemble estimates
techniques. Finally, an ensemble model evaluation was
* * * -
Kumar, Singh, and
Buyya, 2020
Extreme learning
machines
They were using extreme learning machines and a
voting engine to optimize the weights of the prediction
hole-inspired metaheuristic algorithm
* * Heuristics are used to
number of networks
and the number of
hidden nodes in each
network
Zaghloul, Salem, and
Ali-Eldin, 2021
XGBoost algorithm They used real-world datasets to predict query
performance using feature modelling and ensemble
learning based on the duration of time elapsed
* * * -
Jain, 2017 Neural network The event hub and Stream analytics of microsoft azure
with the neural network and self-coded algorithms were
employed to detect fraud telecommunication
* * It cannot be applied
to real-time dynamic
data
Talasila, et al., 2020 Recurrent Neural
Network
rough set and then fed into a recurrent neural network to
predicate the system results
* * -
Zhong, Yu, and Ai,
2020
Deep learning The hierarchical deep learning system used the
behavioural and content features to interpret network
* * *
applied to a small
number of samples
ML: Machine learning
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 53
Abdul-jabbar, S.S. and George, L.E. (2017). Fast text analysis using symbol
enumeration and Hashing methodology. Fast Strings Search Process, 58(1),
pp.345-354.
Abkenar, S.B. Kashani, M.H., Mahdipour, E. and Jameii, S.M. (2020). Big data
analytics meets social media: A systematic review of techniques, open issues,
and future directions. Telematics and Informatics, 57, 101517.
Admiraal, M.M., Ramos, L.A., Delgado Olabarriaga, S., Marquering, H.A.,
Horn, J. and van Rootselaar, A.F. (2021). Quantitative analysis of EEG reactivity
for neurological prognostication after cardiac arrest. Clinical Neurophysiology,
132(9), pp.2240-2247.
Apuke, O.D. (2017). Quantitative research methods : A synopsis approach.
Kuwait Chapter of Arabian Journal of Business and Management Review,
6(11), pp.40-47.
Aung, K.Z. Myo, N.N. (2017). Sentiment Analysis of Students’Comment Using
Lexicon Based Approach. IEEE/ACIS 16th International Conference on Computer
and Information Science, pp.149-154.
Ben Seghier, M., Carvalho, H., Keshtegar, B. and Correia, J.A.F. (2020). Novel
hybridized adaptive neuro-fuzzy inference system models based particle swarm
optimization and genetic algorithms for accurate prediction of stress intensity
factor. FFEMS, 43(11), pp.2653-2667.
Buselli, I., Oneto, L., Dambra, C., Gallego, C.V., Martínez, M.G., Smoker, A. and
Martino, P.R. (2021). Natural Language Processing and Data-Driven Methods for
Aviation Safety and Resilience : From Extant Knowledge to Potential Precursors.
Open Research Europe.
Butcher, B. and Smith, B.J. (2020). Feature engineering and selection: A practical
approach for predictive models. The American Statistician, 74(3), pp.308-309.
Castellanos, C., Pérez, B., Varela, C.A. and Correal, D. (2020). A Model-Driven
Architectural Design Method for Big Data Analytics Applications. Proceedings
2020 IEEE International Conference on Software Architecture Companion,
ICSA-C 2020, pp.89-94.
Cearley, D.W., Natis, Y., Walker, M. and Burke, B. (2018). Top 10 Strategic
Technology Trends for 2018. Gartner, Stamford. Available fromt: https://www.
gartner.com/ngw/globalassets/en/information-technology/documents/top-10-
strategic-technology-trends-for-2018.pdf [Last accessed on 2017 Oct 3].
Chen, M., Mao, S. and Liu, Y. (2014). Big data: A survey. Mobile Networks and
Applications, 19(2), pp.171-209.
Chen, X.W. and Lin, X. (2014). Big data deep learning: Challenges and
perspectives. IEEE Access, 2, pp.514-525.
Choi, B.C., Cho, S. and Kim, C.W. (2018). Kriging Model Based Optimization
of MacPherson Strut Suspension for Minimizing Side Load using Flexible
Multi-Body Dynamics. International Journal of Precision Engineering and
Manufacturing, 19(6), pp. 873-879.
Corizzo, R., Ceci, M. and Malerba, D. (2019). Big Data Analytics and Predictive
Modeling Approaches for the Energy Sector. 2019 IEEE International Congress
on Big Data (BigDataCongress), pp.55-63.
De Fortuny, E.J., Martens, D. and Provost, F. (2013). Predictive modeling with
big data: Is bigger really better? Big Data, 1(4), pp.215-226.
Dean, J. (2014). Big Data, Data Mining, and Machine Learning: Value Creation
for Business Leaders and Practitioners. John Wiley and Sons, Hoboken. Available
from: https://doc.lagout.org/Others/Data Mining/Big Data, Data Mining, and
Machine Learning_Value Creation for Business Leaders and Practitioners
%5BDean 2014-05-27%5D.pdf. [Last accessed on 2022 Apr 01].
Do Nascimento, I.J.B., Marcolino, MS., Abdulazeem, H.M, Weerasekara, I.,
Azzopardi-Muscat, N., Gonçalves, M.A. and Novillo-Ortiz D. (2021). Impact
of big data analytics on people’s health: Overview of systematic reviews and
recommendations for future studies. Journal of Medical Internet Research,
23(4), p.e27275.
Dy, J.G. and Brodley, C.E. (2004). Feature selection for unsupervised learning.
Journal of Machine Learning Research, 5, pp. 848-889.
Elghazel, H. and Aussem, A. (2010). Feature selection for unsupervised learning
using random cluster ensembles. Proceedings IEEE International Conference on
Data Mining ICDM, pp.168-175. Doi: 10.1109/ICDM.2010.137.
Faizan, M. Zuhairi, M.F., Ismail, S. and Sultan, S. (2020). Applications of
Clustering Techniques in Data Mining: A Comparative Study. International
Journal of Advanced Computer Science and Applications, 11(12), pp.146-153.
Fan, C., Xiao, F., Li, Z. and Wang, J. (2018). Unsupervised data analytics
A review. Energy and Buildings, 159, pp.296-308.
Farhan, K.A. and Ali, M.A. (2017). Database Protection System Depend on
2nd International Conference of Cihan University-Erbil
on Communication Engineering and Computer Science, p.2520-4777.
Ghavami P. (2020). Big Data Analytics Methods. 2nd ed. De Gruyter, Berlin.
Hamarashid, H.K., Saeed, S.A. and Rashid, T.A. (2021). Next word prediction
based on the N-gram model for Kurdish Sorani and Kurmanji. Neural Computing
and Applications, 33(9), pp. 4547-4566.
optimization with robustness analysis. Soft Computing A Fusion of Foundations
Methodologies and Applications, 22(19), pp.6371-6394.
Hendrycks, D. Carlini, N., Schulman, J., Steinhardt, J. (2021) Unsolved Problems
in ML Safety. ArXiv, Cornell Tech, pp.1-28. Available from: https://arxiv.org/
abs/2109.13916
Hong, Z., Smart, G., Dawood, M., Kaita, K., Wen, S.W., Gomes, J. and Wu, J.
(2008). Hepatitis C Infection and Survivals of Liver Transplant Patients in
Canada, 1997-2003. Transplantation Proceedings, 40(5), pp.1466-1470.
Hryshchenko, O. and Yaremenko, V. (2021). A comparative analysis of text data
Bayes. Technology Audit and Production Reserves, 5(2(61), pp.6-8.
Jain, V. (2017). Perspective analysis of telecommunication fraud detection
International Journal of Information Technology, 9(3), p.1-8.
Jaouadi, Z., Abbas, T., Morgenthal, G. and Lahemer, T. (2020). Single and
multi-objective shape optimization of streamlined bridge decks. Structural and
Multidisciplinary Optimization, 61(4), pp.1495-1514.
Kan, M.Y. and Klavans, J.L., (2002). Using Librarian Techniques in Automatic
Text Summarization for Information Retrieval. Proceedings of the ACM
International Conference on Digital Libraries, Wuhan, pp.36-45.
Kashyap, R., (2019). Big data analytics challenges and solutions. In: Big Data
Analytics for Intelligent Healthcare Management, Academic Press, Cambridge,
pp.19-41.
Khoshbakht, F., Shiranzaei, A. and Quadri, S.M.K., (2021). Role of the big data
Turkish Journal of Computer and Mathematics Education, 12(10), pp.560-566.
Kumar, D.U., Soon, T.K., Saad, M., Idna, I.M.Y., Mehdi, S. and Bend, H. (2018).
Forecasting of photovoltaic power generation and model optimization : A review.
Renewable and Sustainable Energy Reviews, 81, pp.912-928.
Kumar, J., Singh, A.K. and Buyya, R. (2020). Ensemble learning based predictive
framework for virtual machine resource request prediction. Neurocomputing,
397, p.20-30.
Li, Z., Zhang, Z., Shi, J. and Wu, D., (2019). Prediction of surface roughness in
extrusion-based additive manufacturing with machine learning. Robotics and
Computer Integrated Manufacturing, 57, pp.488-495.
optimization method based on the adaptive approximation model of the radial
basis function. Structural and Multidisciplinary Optimization, 63(4), p.1-19.
Luijken, K., Groenwold, R.H.H., Van Calster, B.E.W., Steyerberg, E.W. and
Van Smeden, M. (2019). Impact of predictor measurement heterogeneity
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
54 http://doi.org/10.14500/aro.10975
across settings on the performance of prediction models: A measurement error
perspective. Statistics in Medicine, 38(18), pp.3444-3459.
Luijken, K., Song, J. and Groenwold, R.H.H., (2022). Quantitative prediction
error analysis to investigate predictive performance under predictor measurement
heterogeneity at model implementation. Diagnostic and Prognostic Research,
1, pp.1-11.
Luijken, K., Wynants, L., van Smeden, M., Van Calster, B., Steyerberg, E.W.
and Groenwold, R.H.H., (2020). Changing predictor measurement procedures
Journal of
Clinical Epidemiology, 119, pp.7-18.
Mariani, M. and Baggio, R. (2022). Big data and analytics in hospitality and
tourism: A systematic literature review. International Journal of Contemporary
Hospitality Management, 34(1), pp.231-278.
McNamara, D.S., Allen, L.K., Crossley, S.A., Dascalu,M. and Perret, C.A.,
(2017). Natural language processing and learning analytics. In: Handbook of
Learning Analytics. Ch. 8. Society for Learning Analytics Research, Alberta,
pp.93-104.
Mishra, R. and Sharma, R. (2014). Big data opportunities and challenges:
Discussions from data analytics perspectives. International Journal of Computer
Science and Mobile Computing, 46(6), pp.27-35.
Moharram, A., Altamimi, S. and Alshammari, R., (2021). Data Analytics and
Predictive Modeling for Appointments No-show at a Tertiary Care Hospital.
2021 1st,
CAIDA 2021, pp.275-277.
Nag, A.K. and Mitra, A., (2002). Forecasting daily foreign exchange rates
using genetically optimized neural networks. Journal of Forecasting, 21(7),
pp.501-511.
Najafabadi, M.M., Villanustre, F., Khoshgoftaar, T.M., Seliya, N., Wald, R. and
Muharemagic, E., (2015). Deep learning applications and challenges in big data
analytics. Journal of Big Data, 2(1), pp.1-21.
Pajouheshnia, R., van Smeden, M., Peelen, L.M., Groenwold, R.H.H., (2019).
transportability of a prediction model. Journal of Clinical Epidemiology, 105,
pp.136-141.
Patel, A., Singh, N.M. and Kazi, F. (2017). Internet of Things and Big Data
Technologies for Next Generation Healthcare. Springer Cham, Berlin.
Pourhomayoun, M., Alshurafa, N., Mortazavi, B., Ghasemzadeh, H., Sideris, K.,
Sadeghi, B., Ong, M., Evangelista, L., Romano, P., Auerbach, A., Kimchi, A. and
Sarrafzadeh, M. (2014). Multiple model analytics for adverse event prediction
in remote health monitoring systems. In: 2014 IEEE Healthcare Innovation
Conference, HIC 2014, pp.106-110.
Rajaraman, V., (2016). Big data analytics. Resonance, 21(8), pp.695-716.
Rajon, S.A.A., Shamim, A. and Arif, M., (2011). A Generic Framework for
Implementing Electronic Commerce in Developing Countries. International
Journal of Computer and Information Technology, 1(2).
Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D.,
Bagul, A., Langlotz, C., Shpanskaya, K., Lungren, M.P. and Na, A.Y., (2017).
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep
Learning. ArXiv.
Rajput, A., (2019). Natural language processing, sentiment analysis, and clinical
analytics. In: Innovation in Health Informatics: A Smart Healthcare Primer,
Academic Press, USA, pp.79-97.
Rashid, T.A., Mustafa, A.M. and Saeed, A.M. (2018). Automatic kurdish text
and Web Technologies, Barolli, L., Zhang, M. and Wang, Z., editors. Lecture
Notes on Data Engineering and Communications Technologies. Vol. 6, Springer,
Cham, Berlin, pp.187-198.
Rocha, J.L.M., Zela, M.A.C., Torres, N.I.V. and Medina, G.S. (2021). Analogy
of the application of clustering and K-means techniques for the approximation
of values of human development indicators. International Journal of Advanced
Computer Science and Applications, 12(9), pp.526-532.
Russell, S. and Norvig, P. (2020). .
4th ed. Prentice Hall, Hoboken.
Saini, N., Saha, S., Chakraborty, D. and Bhattacharyya, B., (2019). Extractive
PLoS One, 14(11), p.e0223477. [Last
accessed on 2022 Apr 01].
Salloum, S., Huang, J.Z., He, Y. and Chen, X. (2018). An asymptotic ensemble
learning framework for big data analysis. IEEE Access, 7(c), pp.3675-3693.
Sanchez-Gomez, J.M., Vega-Rodriguez, M.A. and C., Perez, C.J. (2019). An
Indicator-based Multi-objective optimizationapproach applied to extractive
multi-documenttext summarization. IEEE Latin America Transactions, 17(8),
pp.1291-1299.
Sanchez-Gomez, J.M., Vega-Rodríguez, M.A. and Pérez, C.J. (2018). Extractive
optimization approach. Knowledge-Based Systems, 159, pp.1-8.
Schwarz, C., Schwarz, A. and Black, W.C., (2014). Tutorial: Big data analytics:
Concepts, technologies, and applications. Communications of the Association
for Information Systems, 34(1), pp.1191-1208.
Shamsaldin, A., Rashid, T.A., Fattah, P. and Al-Salihi, N.K., (2019). A study of
the convolutional neural networks applications. UKH Journal of Science and
Engineering, 3(2), pp.31-40.
Shouval, R., Bondi, O., Mishan, H., Shimoni, A., Unger, R. and Nagler, A.,
(2014). Application of machine learning algorithms for clinical predictive
modeling: A data-mining approach in SCT. Bone Marrow Transplantation,
49(3), pp.332-337.
Smith, M., Szongott, C., Henne, B. and von Voigt, G., (2013). Big Data Privacy
Issues in Public Social Media. IEEE International Conference on Digital
Ecosystems and Technologies [Preprint].
Talasila, V., Madhubabu, K., Mahadasyam, M.C., Atchala, N.J. and Kande, L.S.,
(2020). The prediction of diseases using rough set theory with recurrent neural
network in big data analytics. International Journal of Intelligent Engineering
and Systems, 13(5), pp.10-18.
Tran, N.H., Bao, W., Zomaya, A., Nguyen Minh, N.H. and Hong, C.S., (2019).
Federated Learning over Wireless Networks: Optimization Model Design and
Analysis. Proceedings-IEEE INFOCOM, 2019-April(1), pp.1387-1395.
Vargas-Calderón, V., Ochoa, A.M., Nieto, G.Y.C. and Camargo, J.E., (2021).
Machine learning for assessing quality of service in the hospitality sector based
on customer reviews. Information Technology and Tourism, 23(3), pp.351-379.
Verma, J.P. and Agrawal, S., Patel, P. and Patel, A., (2016). Big data analytics:
challenges and applications for text, audio, video, and social media data.
,
5(1), pp.41-51.
Vu, T., Belussi, A., Migliorini, S. and Eldway, A., (2021). Using deep learning
for big spatial data partitioning. ACM Transactions on Spatial Algorithms and
Systems (TSAS), 7(1), p.1-37.
Wang, J., Tang, Y., Nguyen, M. and Altintas, I., (2014). A Scalable Data Science
Proceedings of
the 2014 International Symposium on Big Data Computing, BDC 2014, pp.16-25.
Wang, X.D., Hirsch, C., Kang, S. and Lacor, C., (2011). Multi-objective
optimization of turbomachinery using improved NSGA-II and approximation
model. Computer Methods in Applied Mechanics and Engineering, 200(9-12),
pp.883-895.
Yadav, N. and Chatterjee, N., (2016). Text Summarization using Sentiment Analysis
for DUC Data. International Conference on Information Technology, pp.5.
Yaremenko, V.S., Rogoza, W.S. and Spitkovskyi, V.I., (2021). Application of
Journal of
ARO p-ISSN: 2410-9355, e-ISSN: 2307-549X
http://doi.org/10.14500/aro.10975 55
Theoretical and Applied Information Technology, 99(1), pp.125-134.
Yu, C.H. Lee, H.S., Lara, E. and Gan, S., (2018). The ensemble and model
comparison approaches for big data analytics in social sciences. Practical
Assessment Research and Evaluation, 23(17).
Zaghloul, M., Salem, M. and Ali-Eldin, A., (2021). A new framework based on
features modeling and ensemble learning to predict query performance. PLoS
One, 16(10), pp.1-18.
Zarchi, M. and Attaran, B., (2019). Improved design of an active landing gear
for a passenger aircraft using multi-objective optimization technique. Structural
and Multidisciplinary Optimization, 59(5), pp.1813-1833.
Zheng, L. and Guo, L., (2020). Application of big data technology in insurance
innovation. Journal of Physics Conference Series, 1682(1), pp.285-294.
Zhong, W., Yu, N. and Ai, C., (2020). Applying big data based deep
learning system to intrusion detection. Big Data Mining and Analytices,
3(3), pp.181-195.
Zou, X., et al. (2019). Logistic Regression Model Optimization and Case
Analysis, Proceedings of IEEE 7th International Conference on Computer Science
and Network Technology, ICCSNT 2019, pp.135-139.
