“Stock movement prediction using Machine Learning ”

PROJECT REPORT on
ARTIFICIAL INTELLIGENCE & MACHINE LEARNING
Lab(18CS62)
VI SEMESTER

2021-2022

Submitted by

<Khetan Rishabh Purushotam- 1RV19CS071>

<Mohamed Moin Irfan- 1RV19CS089>

Under the Guidance of

Prof. Revathi S A, Assistant Professor

Department of CSE, RVCE
Bengaluru-560059.
 RV COLLEGE OF ENGINEERING ®, BENGALURU - 560059
(Autonomous Institution Affiliated to VTU, Belagavi)
​
​ DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

CERTIFICATE

Certified that the Lab Project report work titled “Stock movement prediction using
Machine Learning” has been carried out by <Khetan Rishabh
Purushotam(1RV9CS0710 and Mohamed Moin Irfan(1RV19CS089)>, bonafide
students of RV College of Engineering, Bengaluru, have submitted in partial fulfillment
for the Assessment of Course: Artificial Intelligence & Machine Learning (18CS62)
– Lab Component during the year 2021-2022. It is certified that all
corrections/suggestions indicated for the internal assessment have been incorporated in
the report.
 RV COLLEGE OF ENGINEERING, BENGALURU ® - 560059
(Autonomous Institution Affiliated to VTU)

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

DECLARATION

We, <Khetan Rishabh Purushotam(1RV9CS071) and Mohamed Moin

Irfan(1RV19CS089) > the students of 6th Semester B.E., Department of Computer

Science and Engineering, R.V. College of Engineering, Bengaluru hereby declare that

the Lab -project titled “Stock movement prediction using Machine Learning” has

been carried out by us and submitted in partial fulfillment for the Assessment of

Course: Artificial Intelligence & Machine Learning (18CS62) lab component

during the year 2021-2022.

Place: Bengaluru Signature

Date:
 ACKNOWLEDGEMENT
Any achievement, be it scholastic or otherwise does not depend solely on the individual efforts but
on the guidance, encouragement and cooperation of intellectuals, elders and friends. A number of
personalities, in their own capacities have helped me in carrying out this project work. I would like
to take this opportunity to thank them all.

I deeply express my sincere gratitude to my guide Internal Guide, Designation, Department of

CSE, RVCE, Bengaluru, for his able guidance, regular source of encouragement and assistance
throughout this project

I would like to thank Dr.Ramakanth Kumar P, Head of Department, Computer Science &
Engineering, R.V.C.E, Bengaluru, for his valuable suggestions and expert advice.

First and foremost I would like to thank Dr. Subramanya K N, Principal, R.V.C.E, Bengaluru, for
his moral support towards completing my project work.

I thank my Parents, and all the Faculty members of the Department of Computer Science &
Engineering for their constant support and encouragement.

Last, but not the least, I would like to thank my peers and friends who provided me with valuable
suggestions to improve my project.
 ABSTRACT
The stock market is a dynamic and volatile platform which provides an environment and
opportunity for the traders to invest and trade in stocks of particular companies. The price of a
stock is dependent on numerous static and dynamic features. Predicting the trend in future price
movement of a particular company’s stock can be extremely beneficial for investors and traders.
In this project, we predict the direction in which a stock will be moving by studying the
previous trends and using the concept of false positives and training them using the Random
Forest classifier. We have seen a 71% accuracy in our predictions using the given model. The
dataset that has been used for this is the
 Table of Contents

Abstract ii
List of tables
Table 1 - Related works
List of figures

Fig 1 : Dataset Details

Fig 2: Tools and libraries used

Fig 3 data loading

Fig 4 change in price() calculation

Fig 5 NaN Values

Fig 6 Classification Report

Fig 7 Confusion matrix

Fig 8 Feature importance Graph

Fig 9 ROC Curve

1. Introduction
1
1.1. Project Domain and Problem addressing 2
1.2 Issues and Challenges
1.3 Need for AI-based solutions
1.4. Problem Statement
1.5. Project objectives
1.6. Summary
2. Literature Study
(at least 10 to 15 contemporary works presented in running paragraphs,
covering research objectives, methodology, results and gaps). Write the literature
survey in the form of table as shown below:
Sl no Title and year of paper Key findings Drawbacks

Note: Columns could be added if required

3. Design Details
3.1 Architecture
3.2 Methodology
3.3 Data set details
3.4 ML/DL techniques used
3..5 Hardware and Software requirements

4. Implementation details of the Project

4.1 Language/Tools/APIs used
4.2 Data preprocessing
4.3 Validation methodology

5. Results and Analysis

(Detailed discussion on results obtained along with interpretations, evaluation
metrics used, performance analysis with similar works)

6. Conclusion and Future Enhancement

6.1. Novelty in the proposed solution
6.2 Limitations of the Project
6.3. Future Enhancements
6.4. Summary
7. References in IEEE format

Appendices
Appendix A: Screenshots
Appendix B: Printout of the base paper used for implementation of this project
1) Introduction

1.1 Project Domain and Problem addressing

The financial market is a dynamic and composite system where people can buy and sell
currencies, stocks, equities and derivatives over virtual platforms supported by brokers.
Stock markets are affected by many factors causing the uncertainty and high volatility in
the market. Although humans can take orders and submit them to the market, automated
trading systems (ATS) that are operated by the implementation of computer programs can
perform better and with higher momentum in submitting orders than any human. Since
most of the dealings in the markets are done by automated systems, it has now been well
established that training the past data can help us in finding patterns in the movement of the
markets which can be used to predict the future prices. If implemented successfully with a
higher accuracy than existing systems, it could turn into a financial support system with
minimal amount of risk.

1.2 Issues and Challenges

With the amount of volumes involved in the market nowadays, it has become
increasingly difficult to predict stock movement. But using AI and ML systems, we can
forecast assumptions based on previous patterns if we use the right agents to train our
program.

1.3 Need for AI based solutions

AI has a huge potential in the prediction of stock prices. Taking the past performance and
behavior of any stock and training the data available using neural networks and machine
learning models can help in understanding how a stock might behave in the future. Industrially
talking, the system would have huge relevance. It can be used by traders to gain an edge over
others and can also be used by financial institutions for quant-vol trading.
1.4 Problem Statement

With the innovation in technology and their application in the stock market, the system has
become increasingly complex and volatile which in turn has made human predictions
highly inaccurate, but using Machine Learning to find out the patterns in the system using
historical data can help us predict the future prices more accurately.

1.5 Project Objectives

We want to maximize our true positives - days when the algorithm predicts that the price
will go up, and it actually goes up. Therefore, we'll be using precision as our error metric
for our algorithm, which is true positives / (false positives + true positives). This will
ensure that we minimize how much money we lose with false positives (days when we buy
the stock, but the price actually goes down). This means that we will have to accept a lot of
false negatives - days when we predict that the price will go down, but it actually goes up.
This is okay, since we'd rather minimize our potential losses than maximize our potential
gains.

1.6 Summary
It is now evidently clear that AI and ML can have huge significance in topics of prediction
and using these systems in financial markets can be a huge bonus if applied correctly and
carefully. AI systems can predict the movements using knowledge of complex mathematical
functions on the basis of which the stocks move and by training them could be able to predict
how it would move ahead.
 2) Literature Survey

2.1 Introduction:
The work on the use of artificial intelligence and especially machine learning to predict the
prices of any type of equity and commodity has been going on since a long time. With the
increase in the technological developments in the field of Machine learning, it has started
becoming clearer that historical patterns can be used in multiple ways to predict what can
happen in the future relating to the prices of any type of equity or commodity. With this
development, people have started creating more novel models to predict the movements in
prices more accurately. Since these markets are a huge arena for making financial profits, all the
giant financial institutions started conducting even more research in this field to gain an
economic advantage over their competitors and this forced the work on such models to full
force.

2.2 Related Work:

SL. Publications IMPLEMENTATIONS CONS

N
O

1. Neither growth nor pruning

Saurav Agrawal, Dev Artificial neural network methods were attempted for the
Thakkar, Dhruvil Soni, with backpropagation selection of network architecture.
Krunal Bhimani, Dr. algorithm
Chirag Patel, “Stock
Market Prediction using
Machine Learning
Techniques”.
2. Random forest Previous years dataset is
K. Hiba Sadia, Aditya Algorithms, support considered. No real-time data are
Sharma, Adarsh Paul, vector machine used for predicting stocks.
SarmisthaPadhi, Saurav
Sanyal, “Stock Market
Prediction Using Machine
Learning Algorithms”.

3. Root Mean Square Error Doesn't focus on events in the

Murtaza Roondiwala, (RMSE), the difference environment, like news or
Harshal Patel, Shraddha between the target value
and the obtained output social media. It exploits only one
Varma, “Predicting Stock value is reduced by using data source, thus it is highly
Prices Using LSTM”. RMSE value. Recurrent biased.
Neural Network, Long
International Journal of Short-Term Memory
Science and Research
2017.

ISSN: 2319-7064

4. S Abdulsalam Sulaiman Linear regression, Used for limited company stocks

Olaniyi, Adewole, Kayode moving average More amount of data is not
S, considered for prediction
Jimoh, R. G, “Stock Trend
Prediction Using
Regression Analysis – A
Data Mining Approach”.
ARPN Journal of Systems
and Software, Volume 1,
Issue 4, 2011. ISSN:
2222-9833
5. Artificial neural using Bayes theorem bias is
Gareja Pradip, Chitrak network, multiple linear found. Predicted price is
Bari, J. Shiva Nandhini, regression, Bayesian fluctuating they are not
“Stock market prediction Algorithm constant
using machine learning”.

Vivek Kanade, SVM, ANN SVM Only sentiment data are used
6. Bhausaheb Devikar, (Support vector from various news and
Sayali Phadatare, Machine) Twitter resources no historical
Pranali Munde, “Stock data are considered for
market prediction: Using predictions.
historic data analysis”.
International journal of
advanced research in
computer science and
software
engineering, volume 7,
issue 1, 2017. ISSN: 2277
128X. DOI:
10.23956/ijarcsse/V711/01
12.

7. Analyzing the Trend of Random Forest, Support This computational approach

Stock Marketand Evaluate Vector Machine and solves the long outstanding stock
the performance of Market Neural Network market problems and find out the
Prediction using Machine solutions in stock market
prediction research.
Learning Approach by
Mathanprasad L and
Gunasekaran M
8. Stock Market Analysis Support vector Machine This paper is limited to only
using Supervised Machine supervised machine learning,
Learning by Kunal Pahwa and tries to explain only the
and Neha Agarwal fundamentals of this complex
process.

9.
Stock Market Prediction Sentimental analysis in Large amount of storage required
based on Social Sentiments the market for analyzing such sentiments in
using Machine Learning real time.
by Tejas Mankar , Tushar
Hotchandani , Manish
Madhwani , Akshay
Chidrawar, Lifna C.

10. Stock Market Prediction The old statistical The Multi-Layer Perceptron
Using Machine Learning techniques including algorithm of machine learning
Techniques by Mehak Simple Moving predicted 57% correct market
Usmani, Syed Hasan Adil, Average (SMA) and performance.
Kamran Raza and Syed Autoregressive
Saad Azhar Ali
Integrated Moving
Average (ARIMA) are
also used as input. The
machine learning
techniques including
Single Layer
Perceptron (SLP),
Multi-Layer Perceptron
(MLP), Radial Basis
Function (RBF) and
Support Vector
Machine (SVM) are
compared.

Table 1: related works

2.3 summary
The existing system on stock price prediction consists of basic LSTM models and recurrent
neural networks. ANNs use adaptive weights to forecast stock prices. Y. Bing proposed an ANN
to predict the index of the Shanghai Stock Exchange. The authors studied the market between
March 17, 2010 to April 28, 2010. They considered 5 features of the market, open, high, close,
low and volume. The neural network constructed was successful in predicting the daily lowest,
highest, and closing value of the Shanghai Stock Exchange. M. Jia proposed a framework which
made use of the bidirectional long-short term memory (BLSTM) neural network for predicting
the future price of a stock. The authors used the historical data of the GREE stock. They
collected data for 568 days from January 1, 2017 to May 14, 2019. The data consisted of 14
features such as open, high, close, volume etc. The data was normalized and pre-processed.The
close value was used as the benchmark for the prediction. K. A. Althelaya proposed a
Bidirectional LSTM for Short- and Long-Term Stock Market Prediction. The authors had made
use of the Standard and Poor 500 Index (S&P500) historical data for their proposed work.
3) Design Details:

3.1 Architecture
Data is imported to the system using python command. Once the data is obtained, it has to be
cleaned or preprocessed before feeding it to the system to remove any unwanted spikes which
could jitter the results of the prediction. Once the data has been processed, we calculate the
indicators on the basis of which we are going to design our system and train our model. Once all
the requirements have been set we train and test the data by splitting them. Upon testing we can
plot the predicted information through various mathematical tools for better understanding of
the results.

3.2 Methodology

Random Forest is a popular machine learning algorithm that belongs to the supervised learning
technique. It can be used for both Classification and Regression problems in ML. It is based on
the concept of ensemble learning, which is a process of combining multiple classifiers to solve a
complex problem and to improve the performance of the model. As the name suggests,
"Random Forest is a classifier that contains a number of decision trees on various subsets of the
given dataset and takes the average to improve the predictive accuracy of that dataset." Instead
of relying on one decision tree, the random forest takes the prediction from each tree and based
on the majority votes of predictions, and it predicts the final output. The greater number of trees
in the forest leads to higher accuracy and prevents the problem of overfitting.

3.3 Dataset Details

Dataset is imported as csv files to the system. It is basically the [open close high low] of
particular stocks over a period of 10 years or more. The file can be downloaded through official
exchange websites or websites like yahoo finance and kaggle.
 Fig 1 : Dataset Details

3.4 ML/DL Techniques

Random Forests is a supervised machine learning algorithm that uses multiple decision trees in
aggregate to help make more stable and accurate predictions. Decision Trees are the
fundamental building blocks of Random Forest. In essence, Decision Trees is flowlike chart
structure where each node of the tree is used to test a particular attribute of the object. For
example, imagine I have a person which will represent our object. We then test certain attributes
of this person object. For example, one test would be whether they are male or female. The test
will represent a "Decision Node" in our tree, and each of the possible outcomes "Male" or
"Female" will represent a leaf node.The first "Decision Node" in our Decision Tree will be our
"Root Node"

● Root Node: Represents the entire population or sample and this further gets divided into
two or more homogeneous sets. Our starting point.
● Splitting: The process of dividing a node into two or more sub-nodes, for example we
split on gender.
● Decision Node: When a sub-node splits into further sub-nodes, then it is called decision
node.
● Leaf/Terminal Node: Nodes that do not split are called Leaf or Terminal nodes.
● Pruning: When we remove sub-nodes of a decision node, this process is called pruning.
You can say the opposite process of splitting.
● Branch/Subtree: A subsection of the entire tree is called branch or sub-tree.
● Parent and Child Node: A node, which is divided into sub-nodes is called parent node
of sub-nodes whereas sub-nodes are the child of parent node.
In machine learning, we have two categories of learning. Supervised learning and unsupervised
learning. With unsupervised learning, we don't supervise the model and instead allow it to
discover information on its own. We do this by providing an "UNLABELED" data set that
doesn't tell the model what category or value is the "correct" answer. With supervised learning,
we provide the model with a "LABELED" data set which tells the model what the "correct"
value it should be. Random Forest, is an example of a supervised learning algorithm because we
provide the model with a labeled data set.

Reasons for Using random forest method:

1. Instability: Even small changes to the input data can have dramatic changes to the
overall structure of the decision tree.
2. They are often relatively inaccurate. Many other predictors perform better with similar
data.
3. For data including categorical variables with different numbers of levels, information
gain in decision trees is biased in favor of those attributes with more levels.
4. Calculations can get very complex, particularly if many values are uncertain and/or if
many outcomes are linked.

These are some of the reasons it's preferable to use Random Forest because we will see that it
helps overcome some of the weaknesses of Decision Trees.

3.5 Hardware and Software Requirements

Hardware requirements define what sort of hardware specifications we will be working with,
and what will be needed to replicate in some other scenario.

CPU 2 GHz or faster

RAM 4 GB or higher

Disk Space 500 GB SSD or larger

Architecture 32-bit or 64-bit

The specifications given in table x.x is just an estimation. It can vary based on the kind of model
used and the size of the dataset chosen.

Software requirements define what software is being used. It includes major stuff like what kind
of operating system, what databases are being used. The projects’ software requirements are
given in table x.x.

Operating System Windows 10 or newer

Database Obtained through

Kaggle or Yahoo
Finance

Programming Python 3.10.0

(Jupyter notebook)

This project is specifically built in jupyter notebook using python wherein all the dataset
collection (imported through csv file), agents, training and testing of models and the results that
the prediction produces are all implemented using various python libraries like pandas, numpy,
scikit learn etc.
4) Implementation details of the Project

4.1 Language/Tools/APIs used

Project is built in Jupyter Notebook using Python 3.10.0. The tools and APIs that were used
during the implementation were python libraries like numpy, pandas, scikit learn, matplotlib.

Fig 2: Tools and libraries used

4.2 Data preprocessing

Print the price_data data frame to verify the data was loaded.

Fig 3 data loading

With the data now loaded, we can transform so we can calculate some of our trading indicators.
The first thing we need to do is sort the data because we have multiple ticker symbols inside of
our data frame. Take the data frame and call the sort_values method and specify the columns
you wish to sort by using the by argument. In our case, we will be using a list of column names
to sort by. The first sort is by the symbol column, and the second sort is by the datetime column.

Once we've sorted the data, we need to calculate the change in price from one period to the next.
To do this, we will use the diff() method. Grab the close column and call the diff() method. The
diff() method will calculate the difference from one row to the next.

Fig 4 change in price() calculation

Technically, each row where the ticker symbol changes are incorrect because it's using the price
from a different ticker. That means we need to have the first row of each ticker symbol be Nan
for the change_in_price column. To do this, we need to break out into steps.

Step 1: Identify the rows where the ticker symbol changes. If we use the shift() method and shift
every row down by one, the rows where the unshifted column DOES NOT EQUAL the shifted
column is where the ticker changed. We will store these values in a variable called mask.

Step 2: Change those rows to NaN values. We can use the numpy.where() method to test our
series. The test is simple, wherever the mask variable equals True, in other words, wherever the
ticker symbol is different, set the change_in_price column to np.nan.

Fig 5 NaN Values

4.5 Validation methodology:
A total of 6 Indicators are used in the calculation:

1) Relative Strength Index (RSI) : RSI is a popular momentum indicator that determines
whether the stock is overbought or oversold. A stock is said to be overbought when the
demand unjustifiably pushes the price upwards. This condition is generally interpreted as
a sign that the stock is overvalued, and the price is likely to go down. A stock is said to
be oversold when the price goes down sharply to a level below its true value. This is a
result caused due to panic selling. RSI ranges from 0 to 100, and generally, when RSI is
above 70, it may indicate that the stock is overbought and when RSI is below 30, it may
indicate the stock is oversold.

2) Stochastic Oscillator : Stochastic Oscillator follows the speed or the momentum of the
price. As a rule, momentum changes before the price changes. It measures the level of
the closing price relative to the low-high range over a period of time.

3) Williams %R : Williams %R ranges from -100 to 0. When its value is above -20, it
indicates a sell signal and when its value is below -80, it indicates a buy signal.

4) Moving Average Convergence Divergence (MACD) : EMA stands for Exponential

Moving Average. When the MACD goes below the SingalLine, it indicates a sell signal.
When it goes above the SignalLine, it indicates a buy signal.

5) Price Rate Of Change : It measures the most recent change in price with respect to the
price in n days ago.

6) On Balance Volume : On balance volume (OBV) utilizes changes in volume to estimate

changes in stock prices. This technical indicator is used to do buying and selling trends
of a stock, by considering the cumulative volume: it cumulatively adds the volumes on
days when the prices group, and subtracts the volume on the days when prices go down,
compared to the prices of the previous day.
5) Result and Analysis

To get a more detailed overview of how the model performed, we can build a classification
report that will compute the F1_Score, the Precision, the Recall, and the Support. Now, I'm
assuming you don't know what these metrics are, so let's take some time to go over them.

1) Accuracy:
Accuracy measures the portion of all testing samples classified correctly.

2) Recall
Recall (also known as sensitivity) measures the ability of a classifier to correctly identify
positive labels. The recall is intuitively the ability of the classifier to find all the positive
samples. The best value is 1, and the worst value is 0.

3) Specificity
Specificity measures the classifier’s ability to correctly identify negative labels.

4) Precision
Precision measures the proportion of all correctly identified samples in a population of samples
which are classified as positive labels. The precision is intuitively the ability of the classifier not
to label as positive a sample that is negative. The best value is 1, and the worst value is 0.

Fig 6 Classification Report

Plotting of the Confusion matrix:

Fig 7 Confusion matrix

Plotting of Feature Significance Curve:

Fig 8 Feature importance Graph

Plotting of ROC Curve:

Fig 9 ROC Curve

6 Conclusion:

6.1. Novelty in the proposed solution

The novelty in the proposed solution is the use of indicators and not just raw data to train the ML
model. Usage of such indicators help in understanding patterns better as they show market
statistics with a better understanding. We have used a total of six indicators to determine the
agents and all have been explained.

6.2 Limitations of the Project

One major limitation of the project is that it only predicts the direction of the stock and not the
actual prices. If the model would be able to predict prices, then it would be a huge advantage to us
as better trades could be made.

6.3. Future Enhancements

The project has a lot of scope in future and the topic and related research will always be in
demand because the application of this has the power to control the flow of money which will
keep the interest boosted. Hybrid models could be developed to get better accuracy and eliminate
flaws that a single model produced. Apart from that, we can use NLP to view things from a
sentimental analysis point of view.

6.4. Summary
Authors opine that application of machine learning techniques in stock price forecasting needs to
be a well thought process and demands painstakingly detailed execution. The proposed approach
is a paradigm shift in this class of problems by reformulating a traditional forecasting model as a
classification problem. Moreover, knowledge discovery from the analysis should create new
frontiers or applications such as a trading strategy based on the strengths of the classification
accuracy, investigating the behavior of certain classes of stocks.
REFERENCES:

[1] B. Chhimwal and V. Bapat, “Impact of foreign and domestic investment in stock market
volatility: Empirical evidence from India,” Cogent Economics & Finance, vol. 8, no. 1, Apr.
2020.

[2] S. Sridhar, S. Mootha, and S. Subramanian, “Decentralized Stock Exchange Implementation

using Ethereum,” in 2020 International Seminar on Intelligent Technology and Its Applications
(ISITIA), pp. 234- 241, 2020.

[3] C. Pop et al., “Decentralizing the Stock Exchange using Blockchain And Ethereum-based
implementation of the Bucharest Stock Exchange,” in 2018 IEEE 14th International Conference
on Intelligent Computer Communication and Processing (ICCP), pp. 459-466, 2018.

[4] N. Sakthivel and A. Saravanakumar, “Investors’ Satisfaction on Online Share Trading and
Technical Problems Faced by the Investors: A Study in Coimbatore District of Tamilnadu,”
International Journal of Management Studies, vol. V, no. 3(9), p. 71, Jul. 2018.

[5] D. Shah, H. Isah, and F. Zulkernine, “Stock Market Analysis: A Review and Taxonomy of
Prediction Techniques,” International Journal of Financial Studies, vol. 7, no. 2, p. 26, May
2019.

[6] U. Hathi, “Indian Companies Act 2013 Highlights and Review,” SSRN Electronic Journal,
2014.

[7] S. V. Shenoy and K. Srinivasan, “Relationship of IPO Issue Price and Listing Day Returns
with IPO Pricing Parameters,” International Journal of Management Studies, vol. V, no. 4(1), p.
11, Oct. 2018.

[8] G. Tanty and P. K. Patjoshi, “A Study on Stock Market Volatility Pattern of BSE and NSE in
India,” Asian Journal of Management, vol. 7, no. 3, p. 193, 2016.

[9] S. Sridhar, S. Mootha, and S. Subramanian, “Detection of Market Manipulation using

Ensemble Neural Networks,” in 2020 International Conference on Intelligent Systems and
Computer Vision (ISCV), pp. 1–8, 2020

[10] S. Zavadzki, M. Kleina, F. Drozda, and M. Marques, “Computational Intelligence

Techniques Used for Stock Market Prediction: A Systematic Review,” IEEE Latin America
Transactions, vol. 18, no. 04, pp. 744–755, Apr. 2020.

[11] M. C. Joshi, “Factors Affecting Indian Stock Market,” SSRN Electronic Journal, 2013.

[12] S. Alhazbi, A. B. Said and A. Al-Maadid, "Using Deep Learning to Predict Stock
Movements Direction in Emerging Markets: The Case of Qatar Stock Exchange," 2020 IEEE
International Conference on Informatics, IoT, and Enabling Technologies (ICIoT), Doha, Qatar,
pp. 440-444, 2020.
[13] D. Wei, “Prediction of Stock Price Based on LSTM Neural Network,” in 2019 International
Conference on Artificial Intelligence and Advanced Manufacturing (AIAM), pp. 544-547, 2019.

[14] Y. Keneshloo, T. Shi, N. Ramakrishnan, and C. K. Reddy, “Deep Reinforcement Learning

for Sequence-to-Sequence Models,” IEEE Transactions on Neural Networks and Learning
Systems, vol. 31, no. 7, pp. 2469–2489, 2019.

[15] K. Palasundram, N. Mohd Sharef, N. A. Nasharuddin, K. A. Kasmiran, and A. Azman,

“Sequence to Sequence Model Performance for Education Chatbot,” International Journal of
Emerging Technologies in Learning (iJET), vol. 14, no. 24, p. 56, Dec. 2019.

[16] H. Jain and G. Harit, "An Unsupervised Sequence-to-Sequence Autoencoder Based Human
Action Scoring Model," 2019 IEEE Global Conference on Signal and Information Processing
(GlobalSIP), Ottawa, ON, Canada, pp. 1-5, 2019.

[17] K. A. Althelaya, E.-S. M. El-Alfy, and S. Mohammed, “Evaluation of bidirectional LSTM

for short-and long-term stock market prediction,” in 2018 9th International Conference on
Information and Communication Systems (ICICS), pp. 151-156, 2018.

[18] J. Chou and T. Nguyen, "Forward Forecast of Stock Price Using SlidingWindo
Metaheuristic-Optimized Machine-Learning Regression," in IEEE Transactions on Industrial
Informatics, vol. 14, no. 7, pp. 3132- 3142, July 2018.

[19] Y. Zhang and Q. Yang, “An overview of multi-task learning,” National Science Review,
vol. 5, no. 1, pp. 30–43, 2018.