TY Seminar Report - May21

Stock Market Prediction Using Artificial Intelligent & Advanced Architecture 1
Seminar Report
on
Stock Market Prediction Using

Artificial Intelligent & Advanced
Architecture
By
Rishi Acharya
[1032171681]
[PE10]
Under the guidance of
[Name of the Guide]
MIT-World Peace University (MIT-WPU)
MITWPU/SCET/BTECH/Seminar Report
Faculty of Engineering & Technology

School of Computer Engineering & Technology
* 2020-2021 *
MIT-World Peace University (MIT-WPU)
Faculty of Engineering & Technology

CERTIFICATE
This is to certify that Mr. /Ms. Rishi Acharya of B.Tech., School of Computer
Engineering & Technology, Trimester – IX /X, PRN. No. 1032171681, has
successfully completed seminar on
Stock Market Prediction Using Artificial
To my satisfaction and submitted the same during the academic year

20__ - 20__ towards the partial fulfillment of degree of Bachelor of Technology in
School of Computer Engineering & Technology under Dr. Vishwanath Karad
MIT-World Peace University, Pune.
Prof. Rashmi Phalnikar Prof. Dr. Vrushali Kulkarni
Seminar Guide Head

LIST OF FIGURES
2.1. PERFORMANCE OF US STOCK MARKET 2020 2
4.1. DAILY PREDICTION MODEL 5
4.2. FUZZY LOGIC CHART 6
4.3. RECOMMENDATION MODEL 7
2.5. FLOW CHART OF DATA REGRESSION 8
2.6. VOLUME VARIATION 10
2.7. PLAGIARISM REPORT 17
LIST OFTABLES
3.1. LITERATURE REVIEW 3
4.1. EXPERIMENT OUTPUT DATA 11
ACKNOWLEDGEMENT
Apart from the hard work behind the research process of any project , the success of any project
depends largely on the encouragement and guidelines of teachers and mentors. I will take this
opportunity to express my gratitude for my guide Dr. Rashmi Phalnikar ma’am for her all time
coordination and motivations. I would also like to thank her for tremendous support and help. The
guidance and support received from all the friends and research fellows was vital for the success of
this project. Culture provided by the institution made this tough job an opportunity to build
relationships and share ideas for a great outlook.
INDEX
1. UNDERSTANDING THE NEEDS AND CAPACITY OF FINANCIAL
MARKETS...................................................................................................... 1
1.1. FINANCIAL MARKETS................................................................................................ 1

1.2. ROLE IN ECONOMY....................................................................................................1
2. MOTIVATION........................................................................................... 2
2.1. EFFECT OF PANDEMIC ON POPULARITY OF MARKETS............................................. 2
3.LITERATURE REVIEW...........................................................................3
3.1. TABLE OF 3 RESEARCH PAPERS.................................................................................. 3
4. TECHNOLOGY......................................................................................... 4
5. CONCLUSION......................................................................................... 12
6. REFERENCES ........................................................................................ 13
7. CONFERENCES .....................................................................................14
8. PRINT OF BASE PAPER....................................................................... 15
9. PLAGIARAISM ...................................................................................... 17
1. UNDERSTANDING THE NEEDS AND CAPACITY OF

FINANCIAL MARKETS
Financial markets are the backbone of many Political and Profitable Economics. They are generally
considered as temples of capitalism. Financial markets are a part of economies which tell us about
the corporate wealth of the country. The participation of retail and institutional investors in the
game of “buy & sell” creates the momentum of volatility in the price of listed Stocks.
We have some financial products listed and used in Financial markets to make markets keep
running and working. We have few financial products like:
1. Derivatives : Derivatives are investment products that are offered based on the movement of
a specified underlying asset.
2. Bonds : Bonds are one of the most well known fixed income investment products. They can
be offered by governments or corporations looking to raise capital.
3. Stocks : Stock investments represent equity ownership in a publicly traded company.
The stock market is much more random than what most investors think. Indeed, it looks almost like
a random walk: even if it has been going down 7 days in a row, the probability that it will go down
tomorrow is still 50%: random walks are memory-less processes.
2. MOTIVATION
Financial Market refers to a marketplace, where creation and trading of financial assets, such as
shares, debentures, bonds, derivatives, currencies, etc. take place. It plays a crucial role in allocating
limited resources in the country’s economy. The Current Covid-19 situation is changing the nature
of economics as well as ideologies of retail and institutional investors. This brings a high volatility
in Stock Markets. And lack of financial knowledge makes retail investors suffer from huge money
loss. Idea is to create such a tool that can help investors to build their money with better insights.
FIG 2.1. PERFORMANCE OF US STOCK MARKET 2020
3. LITERATURE REVIEW
Table 3.1. Literature review
Author/ Topic Concept Context Future Findings

Date
Xingyu Zhou , Study that Using The major Focusing on GANs, it

1 Zhisong provides Generative parameters has very high
Pan , 1 Guyu prediction Adversarial considered are potential to work
Hu , 1 Siqi methodology Networks to wonders in any real
Opening Price,
Tang,1 and for Stock train models. time forecasting as it
Cheng Zhao1, Markets. Max & Min Contradicts its own
Revised 2018 Price, Trading artificial errors.
volume with
other 9
parameters
Aparna Nayak, Prediction Supervised For Daily Architecture holds an

M. M. Models for machine Analysis unrealistic
Manohara Indian Stock learning We have distribution of tasks
Pai∗ and Market models used to Sentiment in a given time frame.
Radhika M. build model Analysis & The Structuring an
Pai architecture . Historic Data holding of the same it
2016 analysis. necessary
For Monthly
analysis we have
Data driven
algorithmic
analysis.
Ashish Pathak Indian Stock Vast number Sentiment A trading model

and Nisha P. Market of factors Analysis using the proposed
Shetty Prediction influencing the clubbed with methodology can be
2010 Using direction of qualitative & developed to compute
Machine trend on quantitative total returns or
Learning and varying scales analysis. investments in real
Sentiment and multiple time.
Analysis layers.(EMH)
states is used
to determine
quality of
market
4.TECHNOLOGY
4.1. Generative adversarial network (GAN)
It is a framework for estimating generative models via an adversarial process, in which we
simultaneously train two models: a generative model G that captures the data distribution and a
discriminative model D that estimates the probability that a sample came from the training data
rather than G. The training procedure for G is to maximize the probability of D making a mistake.
This framework corresponds to a minimax two-player game. In the space of arbitrary functions G
and D, a unique solution exists, with G recovering the training data distribution and D equal to 0.5
everywhere.
Artificial intelligence with Advanced System Architecture that recruits data from different sources
like news and analyses the market trend accordingly.
Predicting stock prices is not easy work, due to the complexity and chaotic dynamics of the markets
and the many undecidable, nonstationary stochastic variables involved. Many researchers from
different areas have studied the historical patterns of financial time series and have proposed
various methods for forecasting stock prices. In order to achieve promising performance, most of
these ways require careful selection of input variables, establishing predictive models with
professional financial knowledge, and adopting various statistical methods for arbitrage analysis,
which makes it difficult for people outside the financial field to use these methods to predict stock
prices.
4.2. Advanced architecture that comprises various models for decision making for different
aspects for price prediction. Experts suggest that news or viral information plays a big role in price
variation. This follows the technical analysis of charts and a sentiment analysis of News.
Input Data: Yahoo Finance.
We have two models here: one makes the prediction for the next day and another makes the
decision for the upcoming month.
1. All data cleaned and trained by supervised machine learning algorithms.Sentiments are
extracted from the social media and news and then integrated with the parameters to know
affects on price.
2. Technical analysis provides a swing trade mechanism, where the parameters are volumes
and dips.
FIG 4.1. DAILY PREDICTION MODEL
4.3. Fuzzy Logic Module
The main purpose of this module is to create a faithfull list of stocks which are out of manipulation
and have strength of good returns in the long term.
Logic :
IF the News Sentiment was good or the Stock Prediction value was good,
THEN the Stock Faith will be high.
IF the Stock Prediction value was average,
THEN the Stock Faith will be medium.
IF the News Sentiment was poor and the Stock Prediction value was poor,
THEN the Stock Faith will be low.
FIG 4.2. FUZZY LOGIC CHART
4.4. Recommendation Module by Machine Learning
We are using this module to get strength of stock by their Opening and Closing price.
Parameters Included
1. Opening price of prediction day
2. Lowest and highest prices of the prediction day
3. Simple moving average
4. Exponential moving average of opening and closing prices of the prediction day
5. Exponential moving average of lowest and highest prices of the prediction day.
6. Bollinger Bands of opening and closing prices of the prediction day
7. Bollinger Bands of lowest and highest prices of the prediction day
FIG 4.3. RECOMMENDATION MODEL
4.5. Regression Based Model
1. The Regression Models are used for predicting over continuous flow of data by giving them
some independent values.
2. They are considerably used in linear form of functions.
3. V= a + bK + error
4. The factors considered for the regression were low, open, high, close and volume.
5.
6. The R-square confidence test was used to determine the confidence score and the
predictions were plotted to show the results of the stock market prices vs time.
FIG 4.4. FLOW CHART OF DATA REGRESSION
Nomenclature
P Set of positive words
N Set of positive words
U Set of all users
R Set of all user tweets/news
W Set of all words
r Tweet/news of any user
u ∈ U u User u ∈ U
w Any word in the review/news ∈ W
Assigning Polarity.
The Data is moved in a hadoop distributed file. Using a hive query, polarity has been assigned to
each word by comparing existing dictionary.
Volume Variation.
Data recruited from yahoo finance to analyze the trend of the same day to get volume variation
patterns.
Experimental Work
Considering all the research papers I read and understanding the approach of part researcher for this
problem, I tried to come up with my own solution which adds another block in a series of
perspectives.
Task to Accomplish:
1. Real Time sentiment analysis.

2. Filtering of manipulated Stocks
3. Analysing and tracking of Trends as early as possible.
4. TIme-varying Solution.
Considering the above task to cover with one solution, the graph analysis is suggested.
2.6. VOLUME VARIATION
Experiment Data and Results
Stock Selected: Nifty Bank
Time Period : Trading Session of a Full Week
Number of Session = 5
17th May to 21st may
Initial Investment: 20,000
Table 4.1. Experiment data
Date Predicted Price Closing Value Profit/Loss

17 14,900+ 14,929 184+
18 15,000- 15,109 110+
19 16,000+ 15,018 100-
20 15,000- 14,900 55+
21 15,000+ 15,174 189+
After Paying a 2% margin fee to Broker service.
Total Profit Earned Over 5 Trading Session With a Risk of 20,000/- Capital = 438 inr.
5. CONCLUSION
It has been observed that most of the people are investing in the stock market to make money easily.
At the same time investors has high chance of losing all their money. So an efficient predictive
model is required for the user to understand future market trends. There are many predictive models
which tell about the market trend whether it is up or down, but they fail to give accurate results. We
need to build an efficient predictive model of the stock market where the trend for the next day is
predicted. By considering various patterns like continuous up/down, volume traded per day and also
including sentiment of the company a model has been built and tested with different stock market
data available open source. On the considered dataset, Decision Boosted Tree is performing better
than Support Vector Machine and Logistic Regression.
1. 6. REFERENCES
[1] R. Al-Hmouz, W. Pedrycz, and A. Balamash, “Description and prediction of time series: a
general framework of Granular Computing,” Expert Systems with Applications, vol. 42, no. 10,
pp. 4830–4839, 2015.
[2] S. Barak and M. Modarres, “Developing an approach to evaluate stocks by forecasting
effective features with data mining methods,” Expert Systems with Applications, vol. 42, no. 3,
pp. 1325–1339, 2015
[3] A. Booth, E. Gerding, and F. McGroarty, “Automated trading with performance weighted
random forests and seasonality,” Expert Systems with Applications, vol. 41, no. 8, pp. 3651–
3661, 2014
[4] S. A. R. Nai-Fu Chen and Richard Roll, Economic Forces and the Stock Market, The Journal of
Business, vol. 59, no. 3, pp. 383–403, (1986). [Online]. Available
[5] Hellstrom, T. and Holmstrom, K. 1998, ‘Predicting the Stock Market’, Technical Report Series
Ima-TOM-1997-07
[6] Zhisong Pan; hotpzs@hotmail.com Received 6 November 2017; Revised 21 January 2018;
Accepted 13 February 2018; Published 15 April 2018 “Stock Market Prediction on High-
Frequency Data Using Generative Adversarial Nets”
[7] M. Qasem, R. Thulasiram and P. Thulasiram, “Twitter sentiment classification using machine
learning techniques for stock markets,” 2015 International Conference on Advances in
Computing, Communications and Informatics (ICACCI), Kochi, 2015, pp. 834–840.
7. CONFERENCES
1. Published in: 2019 2nd International Conference on Intelligent Computing, Instrumentation

and Control Technologies (ICICICT)
Date of Conference: 5-6 July 2019
Date Added to IEEE Xplore: 13 February 2020
ISBN Information:
INSPEC Accession Number: 19358459
Publisher: IEEE
Conference Location: Kannur, India
2. Published in: 2018 IEEE International Conference on Data Mining Workshops (ICDMW)
Date of Conference: 17-20 Nov. 2018
Date Added to IEEE Xplore: 11 February 2019
INSPEC Accession Number: 18455042
Conference Location: Singapore
8. PRINT OF BASE PAPER
9. PLAGIARISM REPORT
Tool name: https://papersowl.com
FIG 9.1. PLAGIARISM REPORT

TY Seminar Report - May21

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TY Seminar Report - May21

Uploaded by

Copyright:

Available Formats

Stock Market Prediction Using Artificial Intelligent & Advanced Architecture 1

Stock Market Prediction Using

Under the guidance of

[Name of the Guide]

MIT-World Peace University (MIT-WPU)

Faculty of Engineering & Technology

MIT-World Peace University (MIT-WPU)

Faculty of Engineering & Technology

Stock Market Prediction Using Artificial

To my satisfaction and submitted the same during the academic year

Prof. Rashmi Phalnikar Prof. Dr. Vrushali Kulkarni

Seminar Guide Head

4.1. DAILY PREDICTION MODEL 5

4.2. FUZZY LOGIC CHART 6

4.3. RECOMMENDATION MODEL 7

2.5. FLOW CHART OF DATA REGRESSION 8

2.6. VOLUME VARIATION 10

2.7. PLAGIARISM REPORT 17

4.1. EXPERIMENT OUTPUT DATA 11

1.1. FINANCIAL MARKETS................................................................................................ 1

2.1. EFFECT OF PANDEMIC ON POPULARITY OF MARKETS............................................. 2

3.1. TABLE OF 3 RESEARCH PAPERS.................................................................................. 3

8. PRINT OF BASE PAPER....................................................................... 15

1. UNDERSTANDING THE NEEDS AND CAPACITY OF

FIG 2.1. PERFORMANCE OF US STOCK MARKET 2020

Author/ Topic Concept Context Future Findings

Xingyu Zhou , Study that Using The major Focusing on GANs, it

Aparna Nayak, Prediction Supervised For Daily Architecture holds an

Ashish Pathak Indian Stock Vast number Sentiment A trading model

FIG 4.1. DAILY PREDICTION MODEL

4.3. Fuzzy Logic Module

THEN the Stock Faith will be high.

IF the Stock Prediction value was average,

THEN the Stock Faith will be medium.

THEN the Stock Faith will be low.

FIG 4.2. FUZZY LOGIC CHART

4.4. Recommendation Module by Machine Learning

1. Opening price of prediction day

2. Lowest and highest prices of the prediction day

3. Simple moving average

6. Bollinger Bands of opening and closing prices of the prediction day

7. Bollinger Bands of lowest and highest prices of the prediction day

FIG 4.3. RECOMMENDATION MODEL

4.5. Regression Based Model

FIG 4.4. FLOW CHART OF DATA REGRESSION

P Set of positive words

N Set of positive words

U Set of all users

R Set of all user tweets/news

W Set of all words

r Tweet/news of any user

w Any word in the review/news ∈ W

1. Real Time sentiment analysis.

2.6. VOLUME VARIATION

Experiment Data and Results

Stock Selected: Nifty Bank

Time Period : Trading Session of a Full Week

17th May to 21st may

Initial Investment: 20,000

Table 4.1. Experiment data

Date Predicted Price Closing Value Profit/Loss

After Paying a 2% margin fee to Broker service.