Big Data Analytics and Decision Making:

Techniques, Technologies and Challenges

Tanvir Habib Sardar Rashel Sarkar Jameel Ahmed

Dept. of CSE Department of CSE Department of CSE
GITAM School of Technology VNRVJ Institute of Engineering and Alvas Institute of Engineering and
GITAM University, Bangalore, India Technology, Hyderabad, India Technology, Moodbidri, Karnataka
tsardar@gitam.edu jamil.pace@gmail.com

ABSTRACT

The age of big data has come with voluminous, complex, noise and ever fast growing data from
technology driven multiple data sources. Big data encompasses the domains such as healthcare, biomedical,
government, research and commerce. Researchers from diverse fields have been focusing on the big data to obtain
quality knowledge of contribution in their field. The researchers has reported to use many modern tools and
techniques specific to big data in solving the big data decision making task. This paper is focused at (i) concepts
and features associated to big data, (ii) the state-of-the-art techniques and tools for big data decision making, and
(iii) to discover the challenges pertaining to big data so that future researchers can obtain the directions from it.

Keywords—Big data, Decision making; Data analysis; data-intensive processing;

I. INTRODUCTION
The exponential growth of data has arrived from the day-to-day human intractable or passively assistive
tools such as sensors, scientific laboratories, e-commerce or government firms, and social media websites, etc.
The growth of this data has already surpluses the capabilities of microprocessor based processing as the
microprocessor’s growth is slowed down far below to the Moore’s law. A decade below, in 2011, 2.5 quintillion
data bytes were created [1]. It is observed by IBM that 90% of world data is created in each 2 years and exceeding.
Six thousand Walmart stores produces above 267 million of online transactions. In a single day a survey
Telescope’s image data consumes up to 30 trillion bytes [2]. Alibaba produces 20 terabytes of data by over 880
million online transactions. Figure 1 shows the data volume chart as predicted by International data corporation
(IDC) [3]. These data brings us to the definitive conclusion that the era of big data has arrived.

Fig. 1. Global Data Volume Prediction by IDC

Along with the voluminous size of the datasets, big data also defined by the complex data structures and
by a challenging technique of obtaining and managing data [4]. These challenges associated with big data is aimed
at dealing effectively by scientist and engineers, known as Big Data Scientists and Big Data Engineers
respectively. Followed and motivated by the Obama administration in 2012, Gartner put on the big data among
the top ten technology trends and critical technology trends.
The analysis of big data can derive into a big value chain for human life and business profit through the
process of big data identification, integration, and exploitation [5]. The researchers in the big data domain has
consistently developing tools and proposing technological advancement to process the big data and obtain perfect
decision making. The decision making in big data domain has already been experimented successfully in many
domains such as business sale analysis, development of customer specific products, loyalty, patient discovery and
clinical decision making, tourism marketing, and transportation etc. Some survey reported that over half of the
560 surveyed institutions and enterprises have intuition and understanding that big data decision making would
help them in enhancing the efficiency of operation, obtaining the strategic decision, better customer interaction,
advertisement and retention, and setting the product price etc. [6] [2].
Decision support is the key strategy to be included in every sources of big data. The theory of decision
in general means as discovering values, logics, rationalities, probabilistic model deriving and other relative
challenges in the field of economics and computer science using optimal strategy and techniques of mathematics
and statistical modelling. Normative theory of decision making refers to the optimal usage of techniques,
technologies and tools to discover the best decision in a framework of rational bound. Under this definition, the
decision making in big data domain belongs to each steps of big data accumulation to prediction of classified
knowledge system, including the visualization process.
There are many technique and tools developed solely for decision making from big data, although
sometimes they fail to provide optimal solutions. For examples, some of these decision making tools involved
with multi-disciplined areas such as optimization methods, statistics, machine and deep learning, visualization,
data mining, and social media analysis etc. Along with this the big data tools can be divided into three types such
as batch processing, stream processing and hybrid processing tools [2]. Figure 2 explains the relationship between
data science and decision support system (DSS).

Fig. 2. Relationship between Data Science and DSS

The paper is organized as below. The section 2 explains the techniques and tools being used in the big
data processing. The section 3 explains challenges pertaining to big data to get the future direction in research.
The last section concludes the work.
II. TECHNIQUES AND TOOLS OF BIG DATA PROCESSING
A wide variety of tools are developed by scientists and big data enterprises to obtain the valuable
knowledge for their decision making need fulfilment. The tools has crossed multidiscipline domains and proven
effective for discovering, capturing, integrating, analysis and visualizing the big data. This section provide a
comprehensive review of some current trends and development in big data decision making techniques and tools.
The paradigm for big data decision making is shown in figure 3.

Fig. 3. Paradigms of Big Data Processing

A. Techniques of Big Data Decision Making
The techniques of decision making in big data field is either developed or in developing stage due to
enormous complexing in each phase of big data processing. These techniques are usually picked up based on the
specific applications and their respective objectives. There are many techniques which involves many disciplines
and generally overlap with each other. The categories are explained below one by one.

A.A. Mathematical Techniques

The big data study has taken two major concepts from the field of mathematics: statistics and
optimization techniques. The data gathering and analysis or raw data is mainly gone through the statistical
approaches before the actual process of data analysis starts. The numerical description of data points and to derive
the casual relationships or co-relationships between objectives are derived using statistical measures. To deal with
huge volume of big data, some modifications of big data specific statistical techniques are developed such as
parallel statistics [7], statistical computing [8], and statistical learning [9].
Due to effectiveness in quantitative, the optimization techniques is being used in big data. The issues of
high memory cost and time consumption, the optimization techniques are generally incorporated with data
reduction techniques [10] or parallelism [11]. Several studies of co-evolutionary algorithms [12], real-time
optimization [13] in wireless sensor networks [14] and intelligent transportation systems [15] has shown
remarkable use of optimization techniques in big data decision making.

A.B. Data Analysis Techniques

Data analysis takes the advantages of different beneficial techniques for big data, developed under
different domains such as data mining, machine learning, artificial neural networks, and signal processing etc.
Data mining is a set of techniques offered to obtain useful patterns and information from the datasets.
Big data mining requires to improve the existing data mining techniques. For this, generally parallel
implementations [16] of data mining algorithms and dimensionality reduction methods [17] are also considered.
The most widely used data mining techniques used in big data are clustering and regressing. However,
classification and fuzzy logic based data modelling is also reported to be used in the literature.
Machine learning is a sub field of artificial intelligence. Machine learning is classified into supervised
learning and unsupervised learning. In supervised learning the output data labels are known for a certain input set.
A well designed supervised machine learning algorithm learns the big data inputs for the specific output
assignment, as a training part. After the training process, the algorithm can automatically classify the big data
inputs to its respective output class. In unsupervised learning, the data are not labelled. The unsupervised learning
algorithms automatically classify the input big data into different groups based on inherent similar features in the
dataset. The machine learning algorithms are applied successfully in many big data types such as biological big
data [18], sensors data [19], and stock market [19] for example.
Artificial neural network is widely used method of learning useful patterns from the datasets using
distributed node based processing which emulates the human brains. The big data processing using artificial neural
network is however a challenging task due to the artificial neural network’s requirement of many hidden layers
and nodes construction for a well performed learning mechanism. However, more hidden layers and nodes
consumes more time and memory of computing model which has a deep adversarial effect to the big data analysis.
This adversarial effect is however addressed by proposing (i) a sampling method which reduces the sizes of the
big data before input to the artificial neural network and (ii) by designing a parallel or distributed artificial neural
network for big data processing [20]. To contain big data learning, many specific deep learning methods are also
designed by the researchers [21]. It is interesting to know that many deep learning models has successfully dealt
for different big data applications such as drug discovery [22], genomic medicine [23], and text mining [24].
Textual big data, especially from social media and ecommerce websites, is widely used for analysis of
human sentiments. This process of learning involved few machine learning and lexicon-based methods and known
as sentiment analysis [25]. The sentiment analysis is aimed at different applications such as subjectivity
classification [26], polarity determination [27], spam detection [28], review usefulness measurement [29], aspect
extraction [30], and so on. To implement sentiment analysis for the big data, some new big data platforms are
chosen such as MapReduce and Storm [31].

A.C. Visualization Techniques

The visualization technique refers to the ways required for displaying the information using intuitive
display like tables, images, and diagrams. Every big and small firms uses the visualization to display the compact
but meaningful explanation of the datasets and knowledge. For example, Facebook uses timeline as a visualization
method for manipulate and display its data in its dataset. Big data visualization would increase the meaningfulness
but the complexity of big data make it a difficult cases than the traditional datasets [32]. To deal with big data,
many modern visualization techniques trues to discover proper visualization after reducing the dataset size by
feature extraction and parallel way of execution [33] [34]. The best visualization is mandatory for big data as a
good visualization demonstrate a good visualization is better than a thousand of petabytes [35].

A.D. Cloud Computing

Cloud computing is a new revolution in the computing where the required resources and service is rented
to the end users, without requiring to actually buy the resources or the software and installing these to the user
end. The cloud computing model is suitable for big data processing. The decision support can be accomplished
by cloud computing is the data management, tuning of models, data quality, and data currency is properly
maintained [36].

A.E. Fuzzy Sets and Systems

Fuzzy set is an effective solution for many big data problems. Fuzzy set and logic is good for dealing
with uncertainties and vagueness in the dataset, which is very common for big datasets [37]. The fuzzy logic is
thus made to be a good choice for extracting knowledge from incomplete big data [38]. Fuzzy systems are even
used for deep learning in big data analysis and applications as well [39]. The fuzzy techniques being used for big
data are evolving fuzzy systems [40], neural fuzzy classifier [41], linguistic fuzzy rule-based classifier [42], and
fuzzy C-Means is proven best for clustering the big data [43]. Also, many pattern recognition algorithm such as
fuzzy inference systems, fuzzy Bayesian process [44], and fuzzy query system [45] are also experimented for big
data applications. Big data dimensionality reduction can also be performed neural fuzzy classifiers [46].

B. Technologies of Big Data Decision Making

The characteristics and features of big data requires novel technologies, in terms of infrastructures and
platforms, for providing timely and accurate decision making. Figure 4 shows a historical perspective of the big
data framework of technologies. MapReduce changed the batch processing framework and brought a revolution
in batch based big data processing [47], proposed by google in 2003. However, that time only large datasets were
being processed rather than the big data.

Fig. 4. The Three Generations of Processing Paradigms

With the arrival of Hadoop in 2006, the first generation of big data processing started. Hadoop uses
MapReduce as its processing engine. The second generation of big data processing was started by S4 (a Yahoo
product of 2010). S4 dealt with both the static and big data. The hybrid processing can bring us to the third
generation. However, the enough development is this area is yet to happen to let us inter into the third generation.
Table 1 simplifies with its visualization regarding the three generations of technologies by the detailed processing
technologies.
Table 1. Three Generations of Big Data Technologies
Paradigm Technology
Batch Processing MapReduce
Hadoop
Flume
Scribe
 Dryad
Apache Mahout
Jaspersoft BI Suite
Pentaho
Skytree Server
Cascading
Spark
Tableau
Karmasphere
Pig
Sqoop
Stream Processing Kafka
Flume
Kestrel
Strom
S4
SQLstream
Splunk
SAP Hana
Spark Streaming
Hybrid Processing Lambdoop
SummingBird
Batch processing takes care of the data which is stored in storage. The advantages associated with batch
processing are scalability and reliability. The scalability is achieved by parallel implementations like that of
MapReduce. The stream processing on the other hand process big data in real time. This paradigm takes diskless
processing approach to achieve low latency. The hybrid processing synthesizes both the batch and stream
processing based on Lambda architecture [48].
III. CHALLENGES PERTAINING TO BIG DATA
The ultimate target is to develop the big data solutions for decision making which were never before
available. In this section, we shall discuss the challenges in big data decision making and the future solutions to
it. There are many factors which influences the decision making process for big data. The literature reports and
studies show that the factors and its impacts changes over time. The big data and its analysis for decision making
is an ad-hoc process where the organizations changes are frequently altered for obtaining quality output. The
agreements are changed to obtain big data, new staff are hired and new departments are formed so as to obtain
advantages by discover features from big data and subsequent decision making in a short span of time. The factors
which affect the decision making from big data is listed in table 2.

Table 2. Factors Influencing the Decision Making Quality

Factors Description
Contractual The making of agreements and contracts with big data providers is used to increase the data
governance quality. Agreements among organizations are used to ensure mutual understanding of big
data, to create clear responsibilities and procedures, and to improve communication.
Relational Relational governance is necessary for building trust among organizational entities and for
governance ensuring the sharing of relevant knowledge that is necessary to interpret big data. Good
relational governance includes communication and knowledge exchange which is necessary
to understand and process data.
Big data Analyzing big data analytics can contain dozens of variables and parameters. It was difficult
analytics to find the right tools for analyzing. Which techniques can be possibly used and how big data
capabilities can be visualized is a challenge. This was often a long search process in which knowledge of
big data, big data analytics and the domain was necessary
Knowledge Both data and knowledge about the data needs to be transferred. Knowledge about how the
exchange data is collected and processed is necessary for being able to interpret the data and to
understand how it can be used. Once big data analytics analyst have more knowledge about
the context the use of big data analytics and the finding of patterns and relationships becomes
easier
Collaboration The ability to collaborate among big data providers, big data analytics analysts and decision-
makers is a key condition to overcome fragmentation and create a big data chain.
 Furthermore, the inability to collaborate with data providers and to acquire the data can block
the creation of valuable applications
Process The ability to integrate processes and to standardize tasks and data results in enhancing the
integration and big data chain. This results in lower efforts and cost to use big data and big data analytics.
standardization This is important condition for standardizing and routinizing the use of big data.
Routinizing By routinizing big data chain the big data velocity is improved. This helped inspectors to
and make decisions in real-time.
standardization
Flexible Having a flexible infrastructure determines the ability and the amount of effort necessary to
infrastructure handle and process the data. Systems integration improves the handling of big data. Initially
much manual work was necessary which resulted in long lead times for arriving at results.
Staff Finding specialists who can deal with big data, have knowledge of big data analytics, and are
able to communicate with business persons to interpret the results are a key conditions. These
people are scarce. Partnership with companies enabled to use people from outside the Tax
authority.
Data quality of Big data provides little value if it is not accurate and people are not able to interpret the
the big data decisions. Wrong decisions can even be more costly. Wrong decision had even a societal
sources impact, as this resulted in questions asked by politicians
Decision- Decision-makers should be able to interpret the outcomes of the analytics and understand the
maker quality implications. In the case it was found that the more experienced decision-makers were, the
better and the faster decisions could be made.

The main challenges discovered in decision making can relate to the velocity, validity and veracity and
these are connected with the following:
1. Processing: The velocity of data sometimes let the application obtain and deal with just a part of data, leaving
another part behind. This makes the decision making poor as the entire picture of the dataset becomes not so clear.
For example, some part of the data which shows a behavior like fraudulent becomes unknown if that part of data
in unavailable.
2. Noise: The presence of noise creates a problems on data perception and to obtain key insight becomes a problem
in case of noise presence.
3. Error: In many cases the source only has the information on the context of the data. The data analytics have no
idea on data context in this cases. For example, a data may be collected two years back and it reflect the scenarios
of last two years but wrongly it was communicated that the data is of last years. This is an error of data context
and the decision making from this data would be wrong.

IV. CONCLUSION
Big data is a popular domain which is still developing in an enormous speed. The big data decision
making is a new sub-domain of big data analysis which encompasses the techniques and technologies of many
other domains. The techniques and technologies of big data is presented in this paper while providing textual
details and also using few figures and tables also. The paper then provides the challenges associated with big data
processing and discovers factors which influences the big data decision making from the literature.
REFERENCES
[1] Hilbert, Martin, and Priscila López. "The world’s technological capacity to store, communicate, and compute information." science 332,
no. 6025 (2011): 60-65.

[2] Chen, CL Philip, and Chun-Yang Zhang. "Data-intensive applications, challenges, techniques and technologies: A survey on Big
Data." Information sciences 275 (2014): 314-347.

[3] Tien, James M. "Big data: Unleashing information." Journal of Systems Science and Systems Engineering 22, no. 2 (2013): 127-151.

[4] R. Casado et al.Emerging trends and technologies in big data processing, Concurr. Comp-Pract. E., (2015).

[5] Miller, H. Gilbert, and Peter Mork. "From data to decisions: a value chain for big data." It Professional 15, no. 1 (2013): 57-59.

[6] Manyika, James, Michael Chui, Brad Brown, Jacques Bughin, Richard Dobbs, Charles Roxburgh, and Angela Hung Byers. Big data: The
next frontier for innovation, competition, and productivity. McKinsey Global Institute, 2011.

[7] Mahani, Alireza S., and Mansour TA Sharabiani. "SIMD parallel MCMC sampling with applications for big-data Bayesian
analytics." Computational Statistics & Data Analysis 88 (2015): 75-99.

[8] Wilkinson, Leland. "The future of statistical computing." Technometrics 50, no. 4 (2008): 418-435.
[9] Hastie, Trevor, Robert Tibshirani, Jerome H. Friedman, and Jerome H. Friedman. The elements of statistical learning: data mining,
inference, and prediction. Vol. 2. New York: springer, 2009.

[10] Yan, Jun, Ning Liu, Shuicheng Yan, Qiang Yang, Weiguo Fan, Wei Wei, and Zheng Chen. "Trace-oriented feature analysis for large-
scale text data dimension reduction." IEEE Transactions on Knowledge and Data Engineering 23, no. 7 (2010): 1103-1117.

[11] Sahimi, Muhammad, and Hossein Hamzehpour. "Efficient computational strategies for solving global optimization
problems." Computing in Science & Engineering 12, no. 04 (2010): 74-83.

[12]Li, Xiaodong, and Xin Yao. "Cooperatively coevolving particle swarms for large scale optimization." IEEE Transactions on Evolutionary
Computation 16, no. 2 (2011): 210-224.

[13]Sardar, Tanvir Habib, and Zahid Ansari. "An analysis of MapReduce efficiency in document clustering using parallel K-means
algorithm." Future Computing and Informatics Journal 3, no. 2 (2018): 200-209.

[14]Sardar, Tanvir Habib, and Zahid Ansari. "Partition based clustering of large datasets using MapReduce framework: An analysis of recent
themes and directions." Future Computing and Informatics Journal 3, no. 2 (2018): 247-261.

[15]Sardar, Tanvir Habib, and Zahid Ansari. "Detection and confirmation of web robot requests for cleaning the voluminous web log data."
In 2014 International Conference on the IMpact of E-Technology on US (IMPETUS), pp. 13-19. IEEE, 2014.

[16]Sardar, Tanvir Habib, and Zahid Ansari. "An analysis of distributed document clustering using MapReduce based K-means
algorithm." Journal of The Institution of Engineers (India): Series B 101, no. 6 (2020): 641-650.

[17]Ansari, Zahid, Asif Afzal, and Tanvir Habib Sardar. "Data categorization using hadoop MapReduce-based parallel K-means
clustering." Journal of The Institution of Engineers (India): Series B 100, no. 2 (2019): 95-103.

[18]Wen, Zhenshu, Wanwei Zhang, Tao Zeng, and Luonan Chen. "MCentridFS: a tool for identifying module biomarkers for multi-
phenotypes from high-throughput data." Molecular BioSystems 10, no. 11 (2014): 2870-2875.

[19]Wang, Yi, Xinli Jiang, Rongyu Cao, and Xiyang Wang. "Robust indoor human activity recognition using wireless signals." Sensors 15,
no. 7 (2015): 17195-17208.

[20]Nedjah, Nadia, Felipe P. da Silva, Alan O. de Sá, Luiza M. Mourelle, and Diana A. Bonilla. "A massively parallel pipelined reconfigurable
design for M-PLN based neural networks for efficient image classification." Neurocomputing 183 (2016): 39-55.

[21]Arel, Itamar, Derek C. Rose, and Thomas P. Karnowski. "Deep machine learning-a new frontier in artificial intelligence research [research
frontier]." IEEE computational intelligence magazine 5, no. 4 (2010): 13-18.

[22]Preuer, Kristina, Richard PI Lewis, Sepp Hochreiter, Andreas Bender, Krishna C. Bulusu, and Günter Klambauer. "DeepSynergy:
predicting anti-cancer drug synergy with Deep Learning." Bioinformatics 34, no. 9 (2018): 1538-1546.

[23]Sardar, Tanvir Habib, Ahmed Rimaz Faizabadi, and Zahid Ansari. "An evaluation of MapReduce framework in cluster analysis." In 2017
International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICICICT), pp. 110-114. IEEE, 2017.

[24]Sardar, Tanvir Habib, Zahid Ansari, and Amina Khatun. "An evaluation of Hadoop cluster efficiency in document clustering using parallel
K-means." In 2017 IEEE International Conference on Circuits and Systems (ICCS), pp. 17-20. IEEE, 2017.

[25]Sardar, Tanvir Habib, Amina Khatun, and Sahanowaj Khan. "Design of energy aware collection tree protocol in wireless sensor network."
In 2017 IEEE International Conference on Circuits and Systems (ICCS), pp. 12-17. IEEE, 2017.

[26]Ansari, Zahid, Tanvir Habib Sardar, Moksud Alam Mallik, and Naveen D. Chandavarkar. "Data mining in soft computing framework: a
survey." (2002).

[27]Sardar, Tanvir Habib, and Ahmed Rimaz Faizabadi. "Parallelization and analysis of selected numerical algorithms using OpenMP and
Pluto on symmetric multiprocessing machine." Data Technologies and Applications (2019).

[28]Siddiqa, Noor, and Tanvir Habib Sardar. "Multi-Layered Security System Using Cryptography and Steganography." (2019).

[29]Sardar, Tanvir Habib, Zahid Ansari, Naveen D. Chandavarkar, and Amjad Khan. "A Methodology for Detecting Web Robot Requests."

[30]Sardar, Tanvir Habib. "A Methodology in Mobile Networks for Global Roaming." Oriental Journal of Computer Science and
Technology 6, no. 4 (2013): 391-396.

[31]Sardar, T. Habib, Zahid Ansari, and Amjad Khan. "A Methodology for Wireless Intrusion Detection System." International Journal of
Computer Applications 975: 8887.

[32]Staff, C. A. C. M. "Visualizations make big data meaningful." (2014): 19-21.

[33]Bennett, Janine Camille, David Thompson, Joshua Levine, Peer-Timo Bremer, Attila Gyulassy, Valerio Pascucci, and Philippe Pierre
Pebay. Analysis of Large-Scale Scalar Data Using Hixels. No. SAND2011-8450C. Sandia National Lab.(SNL-CA), Livermore, CA (United
States), 2011.
[34]Ahrens, James, Kristi Brislawn, Ken Martin, Berk Geveci, C. Charles Law, and Michael Papka. "Large-scale data visualization using
parallel data streaming." IEEE Computer graphics and Applications 21, no. 4 (2001): 34-41.

[35]Childs, Hank, Berk Geveci, Will Schroeder, Jeremy Meredith, Kenneth Moreland, Christopher Sewell, Torsten Kuhlen, and E. Wes
Bethel. "Research challenges for visualization software." Computer 46, no. 5 (2013): 34-42.

[36]Assunção, Marcos D., Rodrigo N. Calheiros, Silvia Bianchi, Marco AS Netto, and Rajkumar Buyya. "Big Data computing and clouds:
Trends and future directions." Journal of parallel and distributed computing 79 (2015): 3-15.

[37]Morente-Molinera, Juan Antonio, Ignacio J. Pérez, M. Raquel Ureña, and Enrique Herrera-Viedma. "Creating knowledge databases for
storing and sharing people knowledge automatically using group decision making and fuzzy ontologies." Information Sciences 328 (2016):
418-434.

[38]Lin, Chun‐Wei, and Tzung‐Pei Hong. "A survey of fuzzy web mining." Wiley Interdisciplinary Reviews: Data Mining and Knowledge
Discovery 3, no. 3 (2013): 190-199.

[39]Kaburlasos, Vassilis G., and George A. Papakostas. "Learning distributions of image features by interactive fuzzy lattice reasoning in
pattern recognition applications." IEEE Computational Intelligence Magazine 10, no. 3 (2015): 42-51.

[40]Iglesias, José Antonio, Alexandra Tiemblo, Agapito Ledezma, and Araceli Sanchis. "Web news mining in an evolving
framework." Information Fusion 28 (2016): 90-98.

[41]Chang, Hsien-Tsung, Nilamadhab Mishra, and Chung-Chih Lin. "IoT big-data centred knowledge granule analytic and cluster framework
for BI applications: a case base analysis." PloS one 10, no. 11 (2015): e0141980.

[42]López, Victoria, Sara Del Río, José Manuel Benítez, and Francisco Herrera. "Cost-sensitive linguistic fuzzy rule based classification
systems under the MapReduce framework for imbalanced big data." Fuzzy Sets and Systems 258 (2015): 5-38.

[43]Lu, Hua-pu, Zhi-yuan Sun, and Wen-cong Qu. "Big data-driven based real-time traffic flow state identification and prediction." Discrete
Dynamics in Nature and Society 2015 (2015).

[44]Ramachandramurthy, Sivaraman, Srinivasan Subramaniam, and Chandrasekeran Ramasamy. "Distilling big data: refining quality
information in the era of yottabytes." The Scientific World Journal 2015 (2015).

[45]Wang, Hai, Zeshui Xu, Hamido Fujita, and Shousheng Liu. "Towards felicitous decision making: An overview on challenges and trends
of Big Data." Information Sciences 367 (2016): 747-765.

[46]Azar, Ahmad Taher, and Aboul Ella Hassanien. "Dimensionality reduction of medical big data using neural-fuzzy classifier." Soft
computing 19, no. 4 (2015): 1115-1127.

[47]Dean, Jeffrey, and Sanjay Ghemawat. "MapReduce: Simplified data processing on large clusters." (2004).

[48]Janssen, Marijn, Haiko van der Voort, and Agung Wahyudi. "Factors influencing big data decision-making quality." Journal of business
research 70 (2017): 338-345.