Name : Dr.

Bhanu Pratap Singh

Course Name : Online Refresher Course in ICT
(03/07/2023 to 16/07/2023)
Roll Number : 14
Title of the project : AI Systems with Special Reference
to Portfolio Asset Management
 MAJOR RESEARCH PROJECT
ON
AI SYSTEMS WITH SPECIAL REFERENCE TO
PORTFOLIO ASSET MANAGEMENT

PART – A

1. Broad Subject : Inter-Disciplinary (Artificial Intelligence and Finance )

2. Area of Specialization : Artificial Intelligence techniques
With special Reference to Portfolio Asset Management.
3. Duration : Two years (Approximately)
4. Principal Investigator
i. Name : Dr. Bhanu Pratap Singh
ii. Sex : Male
iii. Date of Birth : 10th June 1975
iv. Qualification : Ph.D. (CSE)
v. Designation : Assistant Professor.
vi. Address : AURO University, Surat-394510, Gujarat
5. Co – Investigator(s):
i. Name : Dr. Akhilesh Latoria
ii. Sex : Male
iii. Date of Birth : 19th September 1981.
iv. Qualification : Ph.D.
v. Designation : Assistant Professor.
vi. Address :AURO University, Surat-394510, Gujarat
6. Name of the Institution where the project will be undertaken:
a) Department : School of IT and School of Business
b) University/College : AURO University, Surat-394510, Gujarat
(Please mention the name of affiliating University in case of college)
7. Whether the University/College/Institution is approved under Section 2 (f) and 12 (B)
of the UGC Act? Yes.

8. Teaching and Research Experience of Principal Investigator

 a. Teaching experience : 23 years( Including UG and PG)

b. Research experience : 1) one Student has been registered under

Ph.D.
c. Year of award of Doctoral degree : 2016
d. Title of thesis for doctoral degree : “Development of a technique for route
optimization in Ant based algorithm for MANET” at Mewar University,
Chittorgarh, Rajastan
.
e. Publication ; (Principal Investigator)

1. Paper “The Internet of Things (IoT) Revolution: Applications and Security Challenges” Page
number from b385 to b 387.Journal of Emerging Technologies and Innovative Research (JETIR)
Volume -10, Issue-5, June -2023.
2. Paper “Novel Approach to Integrate and Apply Fuzzy Logic into Data Mining” Page number
from a579 to a 583.
Journal of Emerging Technologies and Innovative Research (JETIR)
Volume -10, Issue-5, June -2023
3. Book titled “Cryptography and Network Sequrity), R K Publications, April 2022
4. Paper “A Semantic Machine Learning Algorithm for Cyber Threat Detection and Monitoring
Security” 2021 3rd International Conference on Advances in Computing, Communication Control
and Networking (ICACCCN), IEEE, May 2022
5. Paper “Proceedings of the 1st International Conference on Smart Innovation, Ergonomics and
Applied Human Factors (SEAHF)”
Page number from 150 to 164., Springer, December 2018
6. Paper “ A Characteristics Study of Ant Colony Optimization Algorithms for Routing Problems”
page number from 149 to 155., International Journal of advanced Research in Computer Science
and Software Engineering.
Volume 3, Issue 3, March 2013, March 2023.
 f. Paper published by (Co- Investigator):
Performance analysis for Load Balancing Algorithms in Cloud Computing environment,
International Conference of Emerging Technologies in IT’22, ISBN: 978-93-5607-271- 8
1. “Hybrid neural synchronization blowfish algorithm for secret key exchange over
Public channels” 1-5 –Received 2 November 2020, Accepted 12 November2020,
Available online 2, January 2021. Today: Proceedings- merging Trends in
Materials Science, Technology and Engineering journal, ISSN-2214-7853.
2. Book Chapter- “BLOOD CANCER CLASSIFICATION WITH GENE
EXPRESSION USING MODIFIED CONVOLUTIONAL NEURAL NETWORK
APPROACH” Accepted –November 2021, CRC Press (Taylor and Francis Book) -
Book Title: Artificial Intelligence in Cyber Physical Systems: Principles and
Applications.

3. “A Novel Approach of Pattern Detection using Fuzzy Logic Implementation in

MATLAB” Volume, 08 || Issue, 11 International Journal of Computational
Engineering Research (IJCER) ISSN (e): 2250 – 3005

4. “A Novel Approach Fuzzy Approach to Pattern Recognition for Large Data

through MATLAB” Volume 5, Issue 9 Journal of Emerging Technologies and
Innovative Research-JETIR ISSN-2349-5162 .

5. “A FUZZY APPROACH TO PATTERN RECOGNITION: A REVIEW” Volume 5,

Issue 9 International Journal for Technological Research in Engineering (IJTRE)
ISSN: 2347 – 4718.

6. “Image Compression analysis and evaluation for various image sequences in

reference to intra prediction” Vol. 3, issue 1, ”International Advanced Research Journal in
Science, Engineering and Technology” (IARJSET) - Peer-Reviewed journal ISSN
(Online):2393- 8021 ISSN (Print): 2394-1588.
 PART B

Title: “ARTIFICIAL INTELLIGENCE TECHNIQUES WITH

SPECIAL REFERENCE TO PORTFOLIO ASSET
MANAGEMENT”

Introduction:

Computerized reasoning (AI) is perhaps the most sizzling subject of present situations on the
grounds that it has upset most businesses as of late, and the monetary administrations area is no
special case. With the coming of fintech, which has a specific accentuation on AI, the area has
encountered an upheaval in a portion of its center practices.
Likely the most impacted region is resource the board, as would be considered normal to
Experience the biggest number of occupation cuts soon (Buchanan 2019). A sizable extent of
resource the board organizations are presently utilizing AI and factual models to run exchanging
and venture stages. The expanded use of AI across a scope of errands in resource the board
requires a more precise assessment of the different strategies and applications required, as well
as the accompanying open doors and difficulties they bring to the area. This study gives a
complete outline of a wide scope of existing also, arising utilizations of AI in resource the
executives, featuring the key subjects of discussion. We center around three significant regions:
portfolio the board, exchanging, and portfolio risk the executives. Portfolio the executives
involves making resource distribution choices to build a portfolio with explicit gamble and
return qualities. Simulated intelligence procedures can add to this cycle by working with
essential examination through quantitative or text based information investigation and
creating novel venture techniques. Computer based intelligence procedures can likewise help
work on the
weaknesses of old style portfolio development procedures. Specifically, AI can deliver better
resource return and hazard gauges and tackle portfolio advancement issues with complex
requirements, yielding portfolios with better out-of-test execution contrasted and conventional
methodologies. Exchanging is one more well-known region for AI applications. Considering the
developing rate and intricacy of exchanges, AI procedures are turning into an fundamental piece
of exchanging practice. An especially appealing element of AI is its capacity to handle a lot of
information to create exchanging signals.
Calculations can be prepared to consequently execute exchanges in light of these signs, which
has led to the business of algorithmic (or algo) exchanging. In expansion, AI procedures can
decrease

exchange costs via naturally breaking down the market and consequently recognizing the best
time, size, and setting for exchanges. Simulated intelligence additionally has huge ramifications
for portfolio risk the board. Since the 2008 worldwide monetary emergency, risk the board and
consistence have been at the very front of resource the board rehearses. With monetary
resources and worldwide Markets turning out to be progressively complicated, conventional
gamble models may as of now not be adequate for risk examination. Simultaneously, AI
procedures that learn Presentation Furthermore, develop by utilizing information can give extra
instruments to observing gamble.
In particular, AI helps risk directors in approving and backtesting risk models. Simulated
intelligence approaches can likewise extricate data all the more productively from different
wellsprings of organized or unstructured information and create more exact conjectures of
insolvency and credit risk, market instability, macroeconomic patterns, monetary emergencies,
etc than customary methods. Moreover, robo-prompting has acquired huge public interest in
Ongoing years. Robo-guides are PC programs that give advanced monetary speculation counsel
in view of numerical guidelines or calculations customized to financial backers' requirements
and inclinations. The fame of robo-consultants comes from their progress in democratizing
speculation warning administrations by making them less expensive and more available to
unsophisticated individual financial backers. Rob advisors are especially alluring to youthful and
educated financial backers, such as Generation Y (recent college grads). Simulated intelligence
is the foundation of normal robo-exhorting calculations, which depend vigorously on the use of
AI across all aspects of resource the executives.
We likewise examine various potential impediments of involving AI in resource

the board. Computer based intelligence models are much of the time misty and complex, making
them challenging for administrators to screen and examine. The models' dependence on and
Aversion to information can present a significant wellspring of hazard. Simulated intelligence
models can be inappropriately prepared because of utilizing low quality or deficient information.
Incapable human oversight could prompt methodical accidents, a failure

to distinguish surmising blunders, and an absence of comprehension of venture practices and

execution attribution by financial backers. Ultimately, whether the advantages
related with AI can legitimize its significant turn of events and execution cost is hazy.
The rest of the piece is coordinated as follows. Segment 2 gives

an outline of patterns in AI and of the most widely recognized AI procedures

utilized in resource the board.

AI Trends:-

As of late, the fame of AI overall — and of AI (ML) explicitly — has flooded in both
practice and the scholarly world. Subsequently, the quantity of examination papers
distributed with the watchwords "man-made brainpower" and "AI" has expanded decisively
in the beyond five years. Artificial intelligence is a more extensive idea than ML, since it
alludes to the general utilization of PCs to copy human mental capacities. ML is really a
subset of AI, in which machines can choose and perform activities in light of previous
encounters. Until now, AI applications in finance generally utilize ML procedures, like
factual learning, and in this way the AI mark

applies just in an exceptionally expansive sense (e.g., Gu, Kelly, and Xiu 2020). Also, an
enormous piece of what is marked as AI (or ML) in finance isn't new yet has existed as
factual or econometric displaying methods for a Long time.

The new promotion about AI can be ascribed to three advancements that are not really
connected with the study of AI itself (Giamouridis 2017).

To start with, PC handling and stockpiling limit have worked on amazingly in the previous
10 years, utilizing some longstanding AI procedures plausible.

Ultimately, AI calculations have been improved and turn out to be generally open,
considering their utilization as a rule without the need for master software engineering
information. This multitude of elements have added to the prevalence of AI and ML as
examination subjects in sociologies.

Despite the fact that AI is a wide field that involves a scope of approaches created over the
long run, the new interest in AI is primarily fixated on ML, which is by a long shot the most
famous AI way to deal with date. ML is worried about utilizing information logically to
adjust the boundaries of measurable, probabilistic, what's more, other registering models. It
basically computerizes one or a few phases of data handling. Albeit a broad rundown of
procedures can achieve this robotization, most ML applications in resource the executives,
and even in finance all the more for the most part, depend on various major (classes of)
methods. These incorporate counterfeit brain organizations (ANNs), bunch examination,
choice trees and arbitrary backwoods, developmental (hereditary) calculations, least
 outright shrinkage and determination administrator (LASSO), support vector machines
(SVMs), and normal language handling (NLP). Supplement a gives a nitty gritty, more
specialized depiction of every one of these strategies.

The interest of scholarly examination in unambiguous AI procedures has consistently

expanded in the beyond twenty years, as outlined by the quantity of distributed papers
regarding the matter. A portion of these procedures, like developmental calculations or
brain organizations, were laid out research themes long before ML acquired prevalence.
Then again, SVMs and NLP have acquired interest all the more as of late. Brain
organization, arbitrary woodland, and NLP procedures have encountered the most keen
expansion in their notice in distributed papers during the beyond five years. Supplement B
gives a more nitty gritty view of the utilization of AI procedures in finance research in light
of examining all functioning papers posted on SSRN. The accompanying areas examine
these strategies

Furthermore, their applications with regards to asset management

.
Review of Literature:

Aggarwal, Charu C., and Chandan K. Reddy, eds. 2014. Data Clustering:

Data Clustering is dividing information into numerous gatherings where every perception
imparts a few qualities to different data of interest inside its gathering. Information bunching
issues have been concentrated on in various fields, and this book endeavors to overcome any
barrier by joining the information acquired in ML and information mining writing. The initial
segment of the book examines information grouping techniques (probabilistic, distance based,
thickness based, matrix based, and phantom grouping, which bunches high-layered and large
information). The second part is dedicated to the use of information grouping in various fields
and circumstances (bunching straight out, text, sight and sound, time-series, organic,
organization, and dubious information). At last, the book centers around the suitability of a
specific grouping strategy through assessment methods accessible from various varieties of
group investigation (visual, supervised, troupe, and approval based arrangements).

Aziz, Saqib, and Michael Dowling. 2019. "AI and AI for Risk Management." In Disrupting Finance:
FinTech and Strategy in the 21st Century, altered by Theo Lynn, John G. Mooney, Pierangelo Rosati, and
Mark Cummins, 33-50. Cham, Switzerland: Palgrave Pivot. https://doi. organization/10.1007/978-3-030-
02330-0_3.
The creators give an outline of the use of ML and AI strategies in the gamble the executives space.
They audit the significant writing furthermore, examine the specific procedures utilized in the space of
credit risk, market risk, functional gamble, and consistence.

Ballings, Michel, Dirk Van lair Poel, Nathalie Hespeels, and Ruben Gryp. 2015. "Assessing Multiple
Classifiers for Stock Price Direction Prediction." Master Systems with Applications 42 (20): 7046-56.
https://doi.org/10.1016/j. eswa.2015.05.013.
In this review, the creators look at troupe ML strategies (Random Forest, Ada Boost, and Kernel Factory)
with single classifiers (Neural Networks, Strategic Regression, Support Vector Machines, K-Nearest
Neighbor) with regard to their capacity to foresee the heading of stock cost developments.
In view of AUC (regions under the collector working trademark bend) as an exhibition measure, the
calculations are positioned as follows: Random Woodlands, Support Vector Machines, Kernel Factory,
AdaBoost, Neural Networks, K-Nearest Neighbor, and Logistic Regression. The review presumes that
troupe models ought not to be ignored in foreseeing stock cost directionsolutions).

Objectives of the study:

1. To study the historical background of the Artificial Intelligence with respect to
portfolio management.
2. To understand the how Artificial intelligence is transforming the asset
management industry by enabling fundamental analysts to research and
extract more information faster so they can uncover accurate investment
insights.

3. To assess the how AI won't replace investment managers, but it could improve
returns

4. To evaluate how might AI affect pensions and savings?

5. To evaluate what are some of the ethical considerations while designing AI
model for portfolio asset management.
6. To generate the patterns and various uses of AI/ML in the investment process.

Methodology:
1. Research Design:

The Descriptive design will be used for the study. The purpose of descriptive
design is to understand the AI techniques can contribute to this process by
 facilitating fundamental analysis through quantitative or textual data analysis
and gen- erating novel investment strategies.

2. Sampling Design:

Sampling and Resampling is a methodology of economically using a data

sample to improve the accuracy and quantify the uncertainty of a
population
parameter. Here we will use Cluster Sampling In a clustered sample, we use
the subgroups of the population as the sampling unit rather than individuals.
The population is divided into subgroups, known as clusters, and a whole
Cluster is randomly selected to be included in the study. This type of sampling
is used when we focus on a specific region or area.

Tools for Data Collection:-

Data for the study will be collected through well-structured data repositories available
with GitHub and Kaggle.
Year wise Plan of work and targets to be achieve:

Sr. No. Research Activity Time Required

1 Finalization of Research proposal & 06 Months
Review of Literature
2 Research Methods adopted
I. Finalization of Tool 03 Months
II. Data Collection 03 Months
III. Data Processing 06 Months
6 Report Writing 03 Months
7 Submission of Research Report 03 Months

Budget Estimate:

Sr. No. Items Expenditure Duration

1 Non recurring: 1,60,000 24 Months
a. Books & Journals

2 Recurring Rs. 5,80,000 24 months

a. Research Project
Fellow @
Rs 20,000 per month
Rs. 80,000 12 Months
b. Hiring Services
Rs. 2,90000 24 Months
c. Field Work &
Travel
(for research
 Investigators &
Project Staff)
3 Contingency 65,000 12 Months
Total 1,66,0000