Accelerating Business Growth with Big Data and Artificial Intelligence

Abstract

Artificial Intelligence (AI) is considered to be the fourth industrial revolution. Arti ficial Intelligence with the hel p of big data has transformed all industries around the worl d. Arti ficial intelligence refers to the simulati on of human or ani mal intelligence in computational systems so that they are programmed to think like intelligent beings and mi mic the actions of intelligent enti ties. Computati onal systems which have programmed intelligence can sol ve different real-worl d problems far more accurately and efficiently than computational systems that are deterministic and hardcoded. Since many problems in business and business analytics cannot be sol ved by deterministic systems, AI plays a major role in tackling problems in the business worl d. Machi ne learning and deep learning which are subsets of the fiel d of AI is wi dely used to sol ve and opti mize many problems in business such as marketing, credit card fraud detection, algorithmic tradi ng, customer service, portfoli o management, product recommendation according to the needs of customers, insurance underwriting. AI and big data have revoluti onized the business worl d and this paper discusses some AI and big data technologies that are currently being used to accelerate business growth.

Full text

Accelerating Business Growth with Big Data and Artificial
Intelligence
Awishkar Ghimire1,# , Surendrabik ram Thapa 1, Avinash Kumar Jha2, Surabhi Adhik ari1 and Ankit Kumar2
1Department of Computer Science and Engineering, Delhi Technological University, India
2Department of Civil Engineering, Delhi Technological University, India
#awishkar.ghimire@gmail.com
ABSTRACT
Artificial Intelligence (AI) is considered to be the
fourth industrial revolution. Artificial Intelligence with
the help of big data has transformed all industries
around t he worl d. Arti ficial intelligence refers to the
simulation of human or animal intelligence in
computational systems s o that they are programmed to
think like intelligent beings an d mimic the actions of
intelligent entities . Computational systems which have
programmed in telligence can solve differe nt real- worl d
problems far more accurately and efficiently than
computational systems that are deter ministic and
hardcoded. Since many proble ms in business and
business analytics cannot be s ol ved by deter ministic
systems, AI plays a major role in t ackling problems in
the business world. Machine learning and dee p learning
which are subsets of the fiel d of AI is wi dely use d to solve
and optimize many problems in business s uch as
marketing, cre dit card fraud detection, algorithmic
trading, customer ser vice, portfolio manage ment,
product rec ommendation accor ding to the needs of
customers, ins urance underwriting. AI an d big dat a
have revoluti onized the business worl d and this paper
discusses some AI and big data technologies that are
currently being us ed to accelerate busines s growth.
Keywords— Artificial Intelligence, Big Data, Busines s
Analytics, Decision Making
1. INTRODUCTION
A business can be defined as an enterprising entity or an
organization cons isting of people and ass ets, that is involved
in professional, commercial or indus trial activities to make a
monetary prof it. A business can be considered as the
backbone of the modern economy. The size o f a business
can range from a s mall indus try that offers its s ervices in a
small town to a very large group of indus tries that is located
in many different countries. Businesses can be owned by a
single person as well as many thousands of people [1].
There are many types and forms of businesses and all of
them have been affected by modern technology especially
the larger corporations. In fact, many large companies like
Google, Facebook, Amazon are at a war of technology.
Technology in today’s modern world is rap idly
improving and is making a huge impact in all sectors of the
modern world [2]. It has impacted on medical diagnos is,
bus iness and many vital sectors [32]. Many businesses in
today’s wor ld have started using modern cutting edge
technology to accelerate their growth and to skyrocket their
profits. Artificial Intelligence, Data Science, Big Data,
Internet Of Things (IoT) have co mpletely transformed
bus iness environments and the way people do business. In
the present day, there is not a single field of work that has
not explored the use cases of AI [27]. The uses of AI and
different computational technologies being used in the
manufacturing industries as well as technology industries
can be seen. It has been well documented that machines can
outperform or at leas t match human be ings in a wide range
of activities including emotion-sens ing, tacit judgement and
automation activities [31]. It has been estimated that
computational technologies could take over as much as 47%
of the current jobs in the world in as little as 10 years [3].
Technology has had a huge impact in the economic world
and the rapid developments in technology in the upcoming
years will have completely change the landscape of the
current economy. So, it is paramount to have an
understanding of how the current cutting edge technologies
are impacting the business world. AI is not only being used
to automate tedious manual work but it has also been
successfully us ed to do jobs that require higher-order
creative thinking such as the works of journalists , attorneys,
lab technicians, para legals etc. [3]. Due to AI and other
computational technologies , many jobs are replaced by
computational technologies, but th is is only for lo w-skilled
jobs such as clerical work. The demand for high skilled jobs
and jobs that require the creation of such computational
systems have increased and is s peedily increasing. Soni et
al. [4] in their paper analy ze 100 AI startups in the world
and describe a very interesting finding. The total
investments in the 100 companies they investigated were
$25.88 million in 2011 which increas ed to $1866.6 million
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 441
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

in 2016. In just 6 years the funding in these companies
increased by a whopping 7112.52%. These figures give an
idea of how artificial intelligence and big data is stirring the
current bus iness scenario.
Fig.1. Bar graph of the total investment every year of the 100 AI st art ups
as analysed by Soni et al. [4]
Artificial intelligence and big data are being used in
various aspects of businesses to improve profits or to speed
up laborious work and thus accelerate bus iness growth.
Many works that would have ta ken thous ands of hours for a
human to complete can be done in a matter of minutes by an
automated system. Many companies like cognizant develop
artificially intelligent systems to improve bus iness processes
and revenue [5]. In one cas e the software company
Cognizant developed a s olution of optimum equip ment
utilization using data analysis for a global mining co mpany
that saved $30 million in the capital due to higher
availability of equipment. Thus can be seen that AI and data
analys is is being heavily used in the modern indus try to cut
costs and optimize profits for the bus iness thus accelerating
growth.
2. RELATED WORKS
The major areas in f inance and business where artificial
intelligence and big data solutions are playing a role a re
loan/ins urance underwriting, fraud detection, customer
service, sentiment/news analysis, algorithmic trading,
portfolio manage ment, marketing, reco mmendation systems
of financial products, advertisement recommendation etc.
Some of the AI techniques and their us es in these areas of
bus ines s are discus sed below.
A. Fraud Detection
Fraud is one of the most substantial crises in today’s
economic world. Business es all around the world lose
billions of dollars due to fraudulent activity. According to
various sources, the total fraud los ses reached $27.85 billion
in 2018 and this figure has been estimated to rise to $40.63
billion in the co ming ten year period [6]. Th is figure is mo re
than the annual GDP o f many third world countries.
Therefore, co mputational systems that are capable of
detecting and catching fraudulent activities is quintessential,
because doing so would rapidly accelerate business and
economic growth. Researchers have devised many systems
that incorporate artificial intelligence and machine learning
to create s uch systems. All fraud detection techniques can
generally be classified into two categories, anomaly
detection and misuse detection [7]. Ano maly detection
learns the trans action behaviour of a certain customer and
any new trans action made by the customer is classified as
normal or anomalous according to past transactions made by
that customer. M isuse detection creates the model using a
labelled data set of a ll the customers, and fraudulent activity
is determined according to the general fraudulent patterns .
Awoyemi et al. [8] co mpare t raditional mach ine learning
models for cred it card fraud detection. A datas et consisting
of 284,807 credit card trans actions fro m European
cardholders was used. Since fraud dataset is highly skewed,
oversampling, undersampling techniques along with no
sampling was also used for better and more accurate
comparison. Their paper co mpared naive Bayes class ifier,
logistic regression classifier and k-nearest neighbour
class ifier [9]. All of these are misuse detection techniques.
The naive Bayes classifier gave 97.92% accuracy, the
logistic regression classifier gave 54.86% accuracy and the
k - neares t neighbour classifier gave 97.69% accuracy.
Logistic regression gives much lower accuracy and this is
apparent because it cannot learn non - linear functions.
Xuan et al. [7] in their paper mention a random forest
algorithm for credit card fraud detection. The dataset used
by them comes from a Chinese e-commerce company.
Random forest is an ensemble learning method in which a
set of decision trees is created, and every tree is made with
an independent data set. The performance of the rando m
forest depends on the strength of each tree and the
correlation between the trees [28]. The higher the strength of
each tree and the lower the correlation between the trees , the
better the performance of the algorithm. Xuan et al use two
different random forest models and compare their
performance on the dataset. The two random fores ts differ in
their bas e classifier i.e individual tree in the model. The first
random fores t us es a s imple decision tree as the base
class ifier and the s econd random forest uses
CART(classification and regress ion) trees as the base
class ifier. The accuracy of the first random forest co mes out
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 442
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

to be 91.96% and the accuracy of the second model co mes
out to be 96.77%.
Randhawa et al. [10] co mpare the performance of about
12 machine learning algorithms on credit card fraud
detection. They use a public credit card transaction datas et.
They also us e ensemble methods which use adaptive
boosting and majority voting for classifying an instance.
They experiment on 7 ensemble methods which use
majority voting. The best accuracy was given by an
ens emble of Neural Network and Na ive Bayes classifier.
This ensemble had an accuracy of 99.941% on the dataset
they used. In the real world, co mpanies generally use
ens emble methods for fraud detection.
B. Algorithmic Trading
Trading is one of the key concepts in economics and
bus iness. Trading can loosely be defined as the buying and
selling of an economic entity such as goods, stocks,
currency etc. Trading is undertaken by individuals and
bus inesses to earn a profit. The trading process generally
consists of 4 components, pre-trade analysis, trading signal
generation, trade e xecution and pos t-trade analysis [11].
Algorith mic trad ing can be loos ely defined as the
automation of any of the co mbination of these steps or all of
these s teps. Artificial Intelligence has completely
trans formed this area of business by automating the trading
process and many trading a lgorithms can c reate profit
without any human intervention. A lgorith mic trad ing us ing
artificial intelligence has had a huge impact on bus ines s and
growth.
Fig. 2. Steps in a complete algorit hmic trading system
Roondiwal et al. [12] propose a learning algorith m that
can be used to predict stock prices . The dataset that they use
is of New York s tock e xchange and the dataset consists of
the data, the open price, the close price, the volume. The us e
LSTM ( Long Short Te rm Me mory) which is a type of
recurrent neural network. LSTMs incorporates a me mory
cell which corresponds to neurons in traditional artificial
neural networks. Thes e memory cells can associate
memories in the input and grasp the structure of data and
hence make the prediction accurate. The LSTM they used
consisted of a sequential input layer after which co mes the 2
LSTM layers and a dense layer which uses ReLU activation
function and then finally the output layer which uses the
linear activation function. Roondiwala et a l conducted
various experiments tweaking different parameters and the
best performing LSTM displayed a n RMSE(Root mean
square error) of 0.00859 which is unbelievably low and
hence proved that A rtificial intelligence can be effect ively
used to predict s tock prices.
Colianni et al. [13] us e sentiment analysis to predict the
prices of bitcoins. They use an open-source API to gather
the tweets with the keyword bitcoin. Then the features are
extracted from the tweet in the form of words and their
counts. This feature then becomes an input to the
class ification algorithm which tells the bitcoin price will
increase or decreas e. The classification algorithms
mentioned in their paper are naive bayes clas sifier and
support vector machines. The SVM gives better accuracy
according to their paper. Their a lgorithm is limited in the
sense that it doesn’t tell what the increasing percentage of
the decrease percentage will be, however, bitcoin price
prediction using sentiment analysis in twitter is an
interesting application of artificial intelligence.
C. Customer Service
Artificial Intelligence and data science solutions have
also been us ed to automate customer service [14]. W ith AI
technology businesses no longer need to hire dedicated
customer service officials. A 24/7 chatbot can answer the
different queries of customers and this practice is s een
wide ly s uccessful in the last few years in different
bus inesses. A I can handle questions of the customers from a
huge knowledge database it has learnt from. This
substantially decreas es the hiring cost of a co mpany, and the
same work can be done with much less money, thus
empowering the growth of the business.
Cui et al. [15] developed a chatbot called s uperagent that
acts as a cus tomer service provider for e-co mmerce websites
like amazon. It is an add on and it can work on any e-
commerce website. The system first crawls the page of the
product that the user searches for and gathers the
information from the page. The system uses state of the art
NLP techniques to interpret the question that the user asks
it. To find informat ion regarding the s ubject matter asked
the system uses deep learning models and recurrent neural
networks and further us es NLP to answer the questions in
English. Super agent system has 3 engines that are
effectively used to answer the questions of the customers.
The first engine is FACT QA that is us ed to answer
questions related to the product informat ion like “what is the
specification of the CPU”. The second engine is FAQ search
Pre-trade
analysis
Post-trade
Analysis
Trade
execution
Trading signal
generation
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 443
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

for cus tomer QA pairs. This engine is used to answer
questions like “does it come with ink” etc. The third engine
is Opinion-Oriented Te xt QA used for reviews. This engine
is used to answer questions like “what do you think o f the
des ign of the cup”.The s ystem is also capable of answering
small talk ques tions such as hello etc. Systems like these can
potentially save millions of dollars for a company.
D. Marketing and Product Recommendation
Artificial Intelligence and big data have had a
considerable impact on marketing as well. Data -driven
marketing strategies are much more e ffective than using
human-based marketing strategies. These days youtube.
Facebook, Ins tagram and almost all s ocial media uses
artificial intelligence and machine learn ing systems to
display individualized and personalized advertisements . A
person that is interes ted in guitar is much more likely to be
shown an advertisement re lated to guitar in s uch platforms.
Marketing is arguably the mos t heavily affected area in
bus iness due to artificial intelligence. The centre piece in
marketing today is customer analytics using big data and the
interpretation of this data using various machine learning
models that produce amazing and profitable insights [16].
Sundsoy et al. [17] experimented on an MNO(mobile
network operator) in Asia compar ing big data-driven
marketing strategies to traditional human-bas ed marketing
strategies. The company often used to send marketing SM S
to selected customers regarding some data packs and hope
that the customer would buy it. The customers were selected
by marketing officials. In their experiment sundsoy et al
divides customers into a control group and treatment group.
The control group was selected by the marketing officials of
the company whereas the treatment group was s elected by
using some data-driven machine learning mode ls. After the
experiment, they found that the conversion rate i.e the
customers who were willing to buy the marketed data pack
was 13 times higher in the treatment group(group selected
using data-driven models) than the control group(group
selected by marketing officials). They tested using different
mach ine learning models such as artificial neural networ ks,
support vector machines, and the final model was a
bootstrapped decision tree. This e xperiment clearly shows
the power of data-driven analyt ics and how it can impact
and accelerate business growth. Product Recommendation
in webs ites is also another form of marketing that is widely
being used in today’s world.
Paradarami et al. [18] propose a hybrid reco mmender
system for recommending business products. Their
recommendation s ystem consists of a deep artificial neura l
network that uses reviews along with content -based features
and collaborative features to create accurate
recommendations. This paper demonstrates that ANN
trained for a specific c lass of bus iness can generalize into
other classes of business as well and recommend products
for other class es very effectively as well. It a lso generalizes
into different types of users as well and hence the s calability
of the recommendation system is dras tically improved. This
approach only requires model para meters to be saved in the
memory unlike memo ry-based recommendation systems,
this significantly decreas ing digital memory footprint and
greatly increasing scalability. Their approach can also be
used to implement real-time reco mmendation systems which
work in an instant due to the speed of their s ys tem.
Fig. 3. General scheme of an AI model
E. Cyber Security
Cybercrime is one of the greatest threats to mankind in
the modern world. It is the most prominent threat to any
modern business in the current day. It has been estimated
that the loss due to cybercrimes is projected to reach $6
trillion by 2021, which is staggeringly huge. Technologies
and strategies to fight cybercrime are para mount in business.
Many business es have gone bankrupt due to cybercrimes.
Effective cybersecurity technology against cybercrimes will
invariably help the growth and sustenance of a bus iness [19].
Artificial intelligence and machine learning have been
successfully used to fight against cybercrimes. Application
of machine learning in cyber c rimes is generally categorized
into three categories, intrusion detection, malware analysis
and phishing detection [20].
Preprocessing data
Building Model
Testing Model
Training Model
Data collection
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 444
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

Kim et a l. [21] us ed Long short term memory ( LSTMs )
which is a type of Recurrent neural network fo r intrusion
detection in cybers ecurity systems. The data set they used to
conduct their e xperiments was KDD Cup 1999 datas et.
4,898,431 network traffics were present in the data s et and
each traffic cons isted of 41 features like IP address etc and
22 attacks are differentiated according to their
characteristics. In their experiment, they co mpared many
different machine learn ing algorithms and LSTM -RNN has
the highest accuracy of 96.93%. Another application of
machine learning in cybersecurity is phishing detection.
Gupta et al. [22] p ropose an artificial neural network
with particle swarm optimization(PSO) to detect phishing
URLs. The PSO techniques are used widely because of its
robustness and short computational time [29]. Moreover,
they also have an easy implementation [30]. In their paper,
they compare back propagation neural network and part
PSO neural network and they find that in their dataset, PSO
TABLE I. PAST WORK IS DON E IN THE FIE LD O F BUSINE SS AN ALY TICS
Name of Author Year of
Publication
Methods used Dataset us ed Use cas e
Paradarami et al. [18] 2017 Deep ANN +
content-bas ed
features +
collaborative
features
Yelp academic
dataset
Product
Recommendation
Awoyemi et al. [8] 2017 K-nearest neighour/
naive bayes / logistic
regression
Credit card
trans action of
european
cardholders datas et
Fraud detection
Xuan et al. [7] 2018 Random forest Credit card
trans action dataset
of chines e e-
commerce company
Fraud detection
Randhawa et al. [10] 2018 Adaboost + majority
voting ,
ANN + NB
Public credit card
transaction datas et
Fraud detection
Roondiwala et al. [12] 2017 LSTM New York stock
exchange dataset
Stock prices
prediction using
time s eries data
Colianni et al. [13] 2015 NLP + SVM/NB Twitter data Bitcoin price
prediction using
sentiment analys is
Cui et al. [15] 2018 NLP + Deep ANN E-commerce data Customer service
Kim et al. [21] 2016 LSTM KDD cup 1999
dataset
Intrus ion detection
in cyber security
Gupta et al. [22] 2017 ANN with PSO UCI repository
archive
Phishing detection
in cyber security
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 445
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

neural network gave an accuracy of 93.93% accuracy when
the number of hidden neurons was 67 and the activation
function used was tansig. The highest accuracy that the
Backpropagation algorithm gave was 96.81% when the
number of hidden neurons was 21 and the activation
function was logs ig. Machine learning is heavily influencing
the field of cybersecurity and many co mpanies are using it
to fight cybercrimes and save billions of dollars.
3. FURTHER APPLICATIONS AND DISCUSS ION
Table I shows the past works that have been done in the
last five years or so. The works show that there are ample
areas where artificial intelligence can be used to grow
bus iness in a lot of ways ranging fro m fraud detections to
product recommendation. The recent use of chatbots in
customer service e xperience was a hugely celebrated
success and it shows that there is a great s cope of AI in
bus iness analytics. Similarly, s mall stock traders and bitcoin
traders also use Business analytics to build predictive
models.
In the above section, various papers on artificial
intelligence and data s cience pertaining to the use cases in
bus iness and finance have been discussed. However, AI is
not jus t limited to the above-mentioned areas of finance.
There are a lot more areas of finance and business in which
artificial intelligence and data science have been
successfully used to make huge progres s in a part icular field.
AI is increas ing the efficiency and productivity across the
entire value chain in a business process, right-front the
mo ment the business model is developed to the moment the
bus iness product goes to the hands of the customer. AI is
revolutionizing the busines s process.
One interesting application of AI in bus iness technology
is automatic retail transactions in supermarkets. Amazon
with its Amazon Go retail stores are using this technology.
There are recognition sens ors of various kinds throughout
the s tore that sense what the customer is picking up int the
store and they automatically charge the customer without
the customer having to checkout through the traditional
process [23]. Another area in wh ich data science and AI has
had immense value is credit scoring. AI and ML algorithms
have been successfully used to give a credit score to
cus tomers given the customer data.
Artificial intelligence has also been successfully us ed in
portfolio manage ment. Port folio manage ment is the robust
decision-making process in which funds are allocated in
different types of inves tment products to reduce the risk of
losing too much money. Jiang et a l. [24] propose a
deterministic deep reinforcement learning method for
solving the portfolio management problem when investing
in cryptocurrencies. Their approach doesn’t rely on any
financial theory hence, it is robust and can be extended to
any other financial mar kets as well. Liu et a l. [25] propose a
one-layer recurrent artificial neural network for solving
linear ps eudoconvex optimizat ion problems with constraints
for dynamic portfolio optimizations to improve profits and
reduce the loss of inves tment opportunities.
A lot of Quantitative Finance is another field of business
in which machine learning has played a huge role.
Spiegeleer et a l. [26] propose various machine learning
models for solving traditional quantitative finance problems
like hedging, curve fitting, derivative pricing. They have
also given a novel idea which used a Bayes ian non-
parametric technique called the GPR. They us e ML models
to fit s ophisticated greek profiles and s ummarise different
implied volatility s urfaces.
According to an article published in the Harvard
bus iness review [3 3], three kinds of artificial intelligence are
taking over the bus iness industry. The first kind is process
automation. This included reading contractual and legal
documents using s tate of the art NLP techniques and
extracting and interpreting it using different intelligent
models, transferring relevant data and information fro m call
centre s ystems and emails to the databases of the company.
The second kind is cognitive ins ights. This kind of system
solves problems like what the customer is more likely to buy,
analyze d ifferent wa rranty data to correctly identify quality
problems or s afety in different manufactured problems.
provide insurance companies w ith detailed and more
accurate modelling. The third kind mentioned in the article
is cognitive engagement. This includes systems like
recommendation systems for health treatment that help us ers
get a customized care plan that takes into account each
user's health history and different data. In that article, the
author’s reviewed 152 different use cases of artificial
intelligence in business systems, among which 47% fell into
the process automation category, 38% fell into the cognitive
insights category and 16% fell into the cognitive
engagement category.
4. CONCLUSION AND FUTURE WORKS
In this paper, s everal applicat ions of mach ine learning,
deep learning and data science in the context of bus iness and
their growth have been reviewed. AI has utterly trans for med
the business world, and many people in today’s world have
been working side by side with artificial intelligen ce
systems to solve complex problems in the world and make a
profit out of the s olution. AI has been s een used in various
areas of the business to accelerate its growth. So me fie lds
which have been discussed in this paper are marketing,
customer service, algorithmic t rading, f raud detection,
cybersecurity, portfolio management, sentiment analysis,
credit scoring.
AI has been used successfully to drive the modern
economic wo rld and it w ill be a key player in s haping the
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 446
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

future. Today’s AI is not even as intelligent as human
beings yet it has had a massive impact in the world. The AI
of tomorrow will undoubtedly change the way the business
and how businesses function is seen. Once AI reaches a
certain level, might be able to see systems that can hire
people into a co mpany. Art ificial Intelligent systems might
someday take over the work of a CEO and might overlook
bus inesses and how they function. In the future, AI systems
might do every job from building a business model to
supplying it into the market rendering human intervention
obs olete.
REFERENCES
[1] S. Prakash et al., "Characteristic of ent erprise collaboration system
and its implementation issues in business management," Internat ion al
Journal of Business Intelligence and Data Mining, vol. 16, no. 1, pp.
49-65 , 2020.
[2] S. Adhikari, S. Thapa, and B. K. Shah, "Oversampling based
Classifiers for Categorization of Radar Returns from the Ionosphere,"
in 2020 International Conference on Electronics and Sustainable
Communication Syst ems (ICESC), 2020: IEEE, pp. 975-978.
[3] S. A. Wright and A. E. Schultz, "The rising t ide of artificial
intelligence and business automation: Developing an ethical
framework," Business Horizons, vol. 6 1, no. 6, pp. 823-832, 2018.
[4] N. Soni, E. Sharma, N. Singh, and A. Kapoor, "Impact of Artificial
Intelligence on Business," in Digit al Innovations, Transformation, and
Societ y Conference 2018 (Digits 2018). pp, 2 018, vol. 1 0.
[5] A. Kumar and M. Janakiraman, "Cognizant T echnology Solutions:
Growth and transformation of its data warehousing and business
intelligence division," Journal of Information Technology Case and
Application Research, vol. 10, no. 3, pp. 56-83, 2008.
[6] S. Rajora et al., "A comp arative study of machin e learning techniques
for credit card fraud detection based on time variance," in 2018 IEEE
Symposium Series on Computational Intelligen ce (SSCI), 2018:
IEEE, pp. 1958-1963.
[7] S. Xuan, G. Liu, Z. Li, L. Zheng, S. Wang, and C. Jiang, "Random
forest for credit card fraud detect ion," in 2018 IEEE 15th
International Conference on Networking, Sensing and Control
(ICNSC), 2018: IEEE, pp. 1-6.
[8] J. O. Awoy emi, A. O. Adetunmbi, and S. A. Oluwadare, "Credit card
fraud detection using machine lear ning t echniques: A comparative
analysis," in 2 017 International Conference on Comput ing
Networking and Informatics (ICCNI), 2017: IEEE, pp. 1-9.
[9] A. Saxena et al., "A review o f clustering t echniques and
developments," Neurocomputing, vol. 2 67, pp. 664-681, 2017.
[10] K. Randhawa, C. K. Loo, M. Seera, C. P. Lim, an d A. K. Nandi,
"Credit card fraud detection using AdaBoost and majorit y vot ing,"
IEEE access, v ol. 6, pp. 1 4277-14284, 2018.
[11] G. Nuti, M. Mirghaemi, P. Treleaven, and C. Yin gsaeree,
"Algorit hmic tradin g," Computer, v ol. 44, n o. 11, pp. 61-69, 2011.
[12] M. Roondiwala, H. Pat el, and S. Varma, "Predict ing st ock prices
using LSTM," International Journal of Science and Research (IJSR),
vol. 6, no. 4, pp. 1754 -1756, 2017.
[13] S. Colian ni, S. Rosales, and M. Signo rotti, "Algorithmic t rading of
cryptocurrency based on T witter sentiment analysis," CS2 29 Project,
pp. 1-5, 2015.
[14] S. T hapa, S. Adhikari, and S. Mishra, "Review of Text
Summarization in Indian Regional Languages," in 2020 Internat ional
Conference on Computing Informat ics & Net works (ICCIN), 2020:
Springer.
[15] L. Cui, S. Huang, F. Wei, C. Tan, C. Duan, and M. Zhou,
"Superagent: A customer service chatbot for e-commerce websites,"
in Proceedings of ACL 2017, System Demonstrations, 2017, pp. 97 -
102 .
[16] S. Erevelles, N. Fukawa, and L. Swayne, "Big Data consumer
analytics and the transformation of marketing," Journal o f business
research , vol. 69 , n o. 2, pp. 897-904, 2016.
[17] P. Sundsøy, J. Bjelland, A. M. Iqbal, and Y.-A. de Montjoye, "Big
data-driven marketing: how machine learnin g outperforms marketers’
gut -feeling," in International Conference on Social Computing,
Behavioral-Cultural Modeling, and Prediction, 2014: Springer, pp.
367-374.
[18] T. K. Paradaram i, N. D. Bast ian, and J. L. Wightman, "A hybrid
recommender system using artificial neural networks," Expert
Syst ems with Applications, vol. 83, pp. 300-313, 2017.
[19] A. Saini, M. S. Gaur, V. Laxmi, an d M. Conti, "Colluding browser
extension attack on user privacy and its implication for web
browsers," Computers & Securit y, vol. 63, pp. 14-28, 2016.
[20] G. Apruzzese, M. Colajanni, L. Ferretti, A. Guido, and M. Marchetti,
"On t he effectiveness of machine and deep learning for cyber
security," in 2018 10th International Conference on Cyber Conflict
(CyCon), 2018: IEEE, pp. 371-390.
[21] J. Kim, J. Kim, H. L. T. Thu, and H. Kim, "Long short term memory
recurrent neural network classifier fo r intrusion detection," in 2016
International Conference on Platform T echnology and Service
(PlatCon), 2016: IEEE, pp. 1-5.
[22] S. Gup ta and A. Singhal, "Ph ishing URL detection by using artificial
neural network with P SO," in 2017 2nd International Conference on
Telecommunication and Networks (TEL-NET), 2017: IEEE, pp. 1-6.
[23] C. Campbell, S. Sands, C. Ferraro, H.-Y. J. T sao, and A.
Mavrommatis, "From data to act ion: How marketers can leverage
AI," Business Horizo ns, vol. 63, no. 2 , pp. 227-243, 2020.
[24] Z. Jian g and J. Liang, "Cryptocurrency portfolio man agement with
deep reinforcement learning," in 2017 Intelligent Systems Conference
(IntelliSys), 2017: IEEE, pp. 905-913.
[25] Q. Liu, Z. Guo, and J. Wang, "A one-lay er recurrent neural network
for const rain ed pseudo convex optimization and its application for
dynamic portfolio optimizat ion," Neural Networks, vol. 26, pp. 99-
109, 2012.
[26] J. De Spiegeleer, D. B. Madan, S. Reyners, and W. Schoutens,
"Machine learnin g for quantitative finance: fast derivative pricing,
hedging and fitting," Quantitative Finance, vo l. 18, no. 10, pp. 1635-
164 3, 2018.
[27] S. Thapa, P. Singh, D. K. Jain, N. Bharill, A. Gupta, and M. Prasad,
“Data-Driven Approach based on Feature Selection Technique for
Early Diagnosis of Alzheimer's Disease”, in 2020 International Joint
Conference on Neural Networks ( IJCNN), 2020: IEEE.
[28] A. K. Jha, S. Adhikari, S. Thapa, A. Kumar, A. Kumar, and S.
Mishra, “Evaluation of Factors Affect ing Compressive St rengt h of
Concrete using Machine Learnin g” in 2020 Advanced Computing an d
Communication Technologies for High P erformance Applications
(ACCTHPA), 2020: IEEE.
[29] N. Bansal, R. Gautam, R. Tiwari, S. Thapa, and A. Singh, “Economic
Load Disp atch Using Intelligent Particle Swarm Optimization”, in
2020 2nd International Conference on Intelligent Computing,
Information and Control Systems (ICICCS 2020), 2020: Springer.
[30] N. Bansal, S. Thapa, S. Adhikari, A. K. Jha, A. Gaba, and A. Jha,
“Novel Exponential Particle Swarm Optimization Technique for
Economic Load Dispatch”, 2nd Intern ational Conference on Inventive
Computation and Information Technologies, 2020: Sp ringer.
[31] S. Thapa, S. Adhikari, U. Naseem, P. Singh, G. Bharathy, and M.
Prasad, “Detecting Alzheimer’s Disease by Exploiting Linguistic
Informat ion from Nepali Transcript”, in 2020 27th International
Conference on Neural Information Processing, 2020: Springer
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 447
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.

[32] S. T hapa, S. Adhikari, A. Ghimire, and A. Aditya, “Feature Select ion
Based T win-Support Vector Machine for the diagnosis of Parkinson’s
Disease”, in 2020 IEEE 8th R10 Humanitarian Technology
Conference (R10-HTC), 2020: IEEE
[33] T. H. Davenport and R. Ronanki, "Artificial intelligen ce for the real
world," Harvard business review, vol. 96, no. 1, pp. 108-116, 201
Proceedings of the Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)
IEEE Xplore Part Number:CFP20OSV-ART; ISBN: 978-1-7281-5464-0
978-1-7281-5464-0/20/$31.00 ©2020 IEEE 448
Authorized licensed use limited to: DELHI TECHNICAL UNIV. Downloaded on March 01,2021 at 10:44:44 UTC from IEEE Xplore. Restrictions apply.