Business Analytics: A Review of Predictive Analysis Techniques

This report provides a comprehensive literature review of predictive analysis techniques within the realm of business analytics. It delves into the core concepts of predictive analysis, emphasizing its role in mitigating future risks through the analysis of historical data. The report explores a range of techniques, including regression analysis (linear, logistic, and polynomial) and machine learning methods (neural networks, multilayer perceptron, and radial basis functions). It highlights the significance of data collection, processing, and modeling in the predictive analysis process. The report also examines the impacts of predictive analysis techniques on organizations, such as improved decision-making, cost reduction, and enhanced operational efficiency. Furthermore, the report discusses the application of these techniques across various industries, including banking and healthcare, to address challenges and optimize outcomes. The report concludes by summarizing the key findings and emphasizing the importance of predictive analysis in modern business practices.

Techniques of Predictive Analysis
Abstract— The main purpose of this report is to review the techniques
of predictive analysis in business analytics. In this report several
journals are analyzed for literature review. The main objective of this
report is to provide a brief information of the predictive analysis and
its techniques.
Keywords— Predictive analysis, Regression, Data analysis, Business
analytics, Linear regression, Analysis
I. INTRODUCTION
In Business analytics, predictive analysis plays a very important
role. Predictive analysis provides an opportunity to the
organizations to avoid future risks by analyzing previous data.
There are many different techniques are present in predictive
analysis. Predictive analysis uses highly diverse arsenal of
techniques which helps the organizations. Basically the word
predictive analysis comprises with two words predict and analysis.
According to the order of words predictive analysis works
reversely, that is analyzed the data first, then based on the analysis
made the predictions. Basically predictive analysis deals with
previously perceive data to predict future events by applying some
methods like machine learning. Data is collected from different
types of sources and after collecting the data, some techniques are
applied to transform the data into a well-structured format.
Filtering, data correlating and many other techniques are used to
transform the data.
II. LITERATURE REVIEW
A. Predictive Analysis
According to [1], [2], Predictive analysis is a type of Business
analytics also known as BA. Predictive analytics is a form of
advanced analytics, which uses both new and previous data to
predict activity and trends of the organizations. The predictive
analytics is mainly depends on data and the data is the main aspect
of this analytics because without any data set it is not possible to
predict anything. Data can be collected from the various sources
and filtering and other techniques are used to transform the data
into a structured form because some time in the data set some
values are missing which is reducing the accuracy of prediction.
After that, data is stored in the data warehouse.
According to [1], Predictive analytics is a field of statistics and
different statistical techniques are used like data mining, machine
learning, data modelling, deep learning algorithms and so on. This
can be applied on the event which are unknown, whether it is
present, past or future event. For example, identifying an accused
after a crime or credit card fraud as it happens. Predictive analytics
core is mainly depends on the explanatory variables and predicted
variables from the previous occurrence which is used for
prediction.
Fig-1: Value chain of predictive analytics
Data mining tools and techniques are used for building the
predictive analytics model. The first step involves obtaining data
from the database. Then with the help of advanced algorithms the
data is processed to detect predictive information and hidden
patterns. Although one is a clear relationship between data mining
and statistics, the methodologies used in data mining and data
originated in areas other than statistics.
B. Process of Predictive analytics
Predictive analytics comprise in seven different processes and
the process are Project definition, data collection, data analysis,
statistics, modelling, deployment and model monitoring. Every
process of the predictive analytics is an important process. Process
of predictive analytics is described below.
i. Project Definition: This is the first process of predictive
analytics and in this process the objective of the
business, project outcomes and the data sets are
identified which is used for further processes.
ii. Data Collection: In this process data are collected from
the various sources and data mining techniques are used
in the data sets for the analysis. This process provides
an entire customer interaction view.[6]
iii. Data analysis: In this process the data are processed to
check the data is correct or not its means that the data
cleaning and modelling is done for finding useful
information.
iv. Statistics: Statistic analysis allows to prove the
presumptions, theories and examine those using
standard statistical models.
v. Modelling: In this process of predictive analytics
provides the ability to create predictive models
automatically.

vi. Deployment: This process of predictive analytics
provides possibilities to deploy analytical results into
day to day decisions.
vii. Model Monitoring: This is the last process of the
predictive analytics, which monitors all the models of
predictive analytics to check the models are providing
expected results or not.
Fig-2: Predictive Analytics Process Flow
III. TECHNIQUES OF PREDICTIVE ANALYSIS
The techniques of predictive analytics can be categorized into
two techniques. The techniques of predictive analytics are
regression techniques/analysis and machine learning techniques.
A. Regression techniques
According to [3], Regression analysis is a form of predictive
modelling technique which helps to identify the relationship
between a dependent variable which is also known as target and
independent variable which is also known as a predictor. It is a
very important tool for analyzing and modelling the data. The
regression techniques are used for predicting, modelling time series
and finding the relation between variables. Suppose the relation
between rash driving and the numbers of accidents by drivers is the
pre-eminent studied through regression.
There are many different types of regression techniques which is
used for making predictions they are linear regression, logistic
regression and polynomial regression.
The linear regression technique is one of the most widely known
technique for modelling. In this technique of regression the
relationship of dependent variable and predator variables are
analysed. [7]
According to [8], Logistic regression is also used to describe the
data and also explain the relationship between one dependent
variable which is a binary variable and one or more independent
variables.
Polynomial regression is used to fit data as an estimate to a non-
linear model.
B. Machine learning Techniques
According to [4], Machine learning is a part of artificial
intelligence. Machine learning was initially developed for the
computers to develop the ability to learn like a human. Today, as it
includes many advance statistical methods for classification and
regression. Now, it finds applications in diverse field, including
speech recognition, credit card fraud, stock market analysis and
medical diagnostics. It is enough to predict the dependent variables
without concentrating on the underlying relationships between
variables in some application. But in some other cases the
relationship may be very complex and mathematically unknown of
dependencies. Some of the methods used in predictive analysis are
neural network, MLA that is multilayer perceptron, radial basis
function (RBF).
The neural network is a nonlinear modelling technique capable
of modelling complex functions. [5]
According to [9], Multilayer perceptron contains input and
output layers of sigmoid nodes with some hidden layers.
The Radial basis functions are used for rectifying data. This
function is applied to the neural network areas and this function is
also used to replace sigmoidal transfer functions.
The machine learning technique of predictive analysis used to
deal with big data because now days the numbers of data is
increased due to more data is stored in the data warehouse than
ever before, and the data analysis became more complex. So this
technique of predictive analysis provides a better view of previous
data which helps to analysis, historical data and helps to predict
future events. [4], [10]
IV. IMPACTS OF PREDICTIVE ANALYSIS TECHNIQUE’S
According to [1], [11], Predictive analysis techniques are
playing very vital role in the organizations to get the desired goal
by analyzing historical and new data. Predictive analysis
techniques give an overview of future events and helps the
organizations to deal with future risks of the business.
Organizations are using such techniques to increase productivity
by analyzing old data which are very easily used for analysis
purpose. It is also allowing the organization to be more active and
flourish in the future predictions based on past data and not
traditional estimates.
Predictive analysis also helps the organization to look into the
cost which they are spending on the business for the betterment of
the organization. Sometime organizations are spending a high
amount in a particular field which comes with an additional cost.
Predictive analysis techniques help to reduce the additional cost of
analyzing previous data.
Predictive analysis helps to identify misallocated resources or
additional resource allocation. It also helps to achieve desired
results with the accuracy and saves on any incorrect resource
allocations which further assists to save both time and cost.
It helps to get the fastest result depends on the predictions by
analyzing old data. This technique for prediction analysis takes the
chance to gain advantage of the future trends of the business.
The main impact of this technique of predictive analysis in
business analytics is that, it helps to reduce risk of a future event
before it going to happen by analyzing previous data. Predictive
analysis techniques help to improve operations of the organization
like quality and functionality.
Such techniques are used by several of industries and
organizations like airlines, hospitality and banks to deal with future
events and risk in their decision making process. The bank uses this
technique to deal with cyber threats like credit card and internet
banking fraud with the help of predictive analysis in the real-time
context. [12]

Fig-3: Business value of analytics with respect to time
V. CONCLUSIONS
The main aim of this report is to understand the techniques of
predictive analysis. Predictive analysis helps the organizations to
deal with future risks which are going to happen in the future by
analyzing data. This report is used to give a brief literature review
of the techniques of predictive analysis. In this report the
importance of predictive analysis and why this technique is used in
the current time and some other aspects are discussed. In this
literature review report the process of predictive analytics and the
techniques are described. In these reports, different techniques of
predictive analysis are used in this literature review, and also
discussed why such techniques are important in predictive analysis.
In the above report, the main focus is on the impacts of predictive
analysis techniques in the businesses and why it's necessary.
REFERENCES
[1] J.G. Nies, J. Watson, O. Stern, M. McCreesh, Verint Americas Inc, Predictive
Modeling from Customer Interaction Analysis, U.S. Patent Application
14/083,772, 2014.
[2] C. Holsappl, A. Lee-Post, R. Pakath, A unified foundation for business
analytics. Decision Support Systems, 64, pp.130-141, 2014.
[3] F. Kaytez, M.C. Taplamacioglu, E. Cam, F. Hardalac, Forecasting electricity
consumption: A comparison of regression analysis, neural networks and least
squares support vector machines. International Journal of Electrical Power &
Energy Systems, 67, pp.431-438, 2015.
[4] M.I. Jordan,T.M. Mitchell,Machine learning: Trends, perspectives, and
prospects. Science, 349(6245), pp.255-260, 2015.
[5] S. Han, J. Pool, J. Tran, W. Dally, Learning both weights and connections for
efficient neural network. In Advances in neural information processing
systems (pp. 1135-1143), 2015.
[6] S.M.M. Rubiano, J.A.D. Garcia, Analysis of data mining techniques for
constructing a predictive model for academic performance. IEEE Latin
America Transactions, 14(6), pp.2783-2788, 2016.
[7] L.Yang, S. Liu, S. Tsoka, L.G. Papageorgiou, Mathematical programming for
piecewise linear regression analysis. Expert systems with applications, 44,
pp.156-167, 2016.
[8] P.C. Austin, J. Merlo, Intermediate and advanced topics in multilevel logistic
regression analysis. Statistics in medicine, 36(20), pp.3257-3277, 2017.
[9] H. Ramchoun, M.A.J. Idrissi, Y. Ghanou, M. Ettaouil, Multilayer Perceptron:
Architecture Optimization and Training. IJIMAI, 4(1), pp.26-30, 2016.
[10] A.G. Shoro, T.R. Soomro, Big data analysis: Apache spark perspective,
Global Journal of Computer Science and Technology, 2015.
[11] D. Appelbaum, A. Kogan, M. Vasarhelyi, Z. Yan, Impact of business
analytics and enterprise systems on managerial accounting. International
Journal of Accounting Information Systems, 25, pp.29-44, 2017.
[12] Z. Sun, L. Sun, K. Strang, Big data analytics services for enhancing business
intelligence, Journal of Computer Information Systems, 58(2), pp.162-169,
2018.