STOCK PRICE PREDICTION

A Project Report submitted in partial fulfilment of the requirements

for the award of the degree of

Bachelor of Technology
in
Computer Science and Engineering
by
Ashutosh Sarkar(201500169)
Navneet Singh (201500433)
Himanshu Gupta(201500294)
Amit Kumar Yadav(201500081)

Group No. 182

Under the Guidance of
Mr. Vikash Sawan , Assistant Professor

Department of Computer Engineering & Applications

Institute of Engineering & Technology

GLA University
Mathura- 281406, INDIA
April, 2024
 Department of Computer Engineering and Applications
GLA University, 17 km Stone, NH#2, Mathura-Delhi Road,
P.O. Chaumuhan, Mathura-281406 (U.P.)

Declaration
I hereby declare that the work which is being presented in the B.Tech. Project
“Stock price prediction using machine learning”, in partial fulfillment of the
requirements for the award of the Bachelor of Technology in Computer
Science and Engineering and submitted to the Department of Computer
Engineering and Applications of GLA University, Mathura, is an authentic
record of my own work carried under the supervision of Mr. Vikash Sawan ,
Assistant Professor.
The contents of this project report, in full or in parts, have not been
submitted to any other Institute or University for the award of any degree.

Sign Sign
Ashutosh Sarkar Navneet Singh
201500169 201500433

Sign Sign
Himanshu Gupta Amit Kumar Yadav
201500294 201500081
 Department of Computer Engineering and Applications
GLA University, 17 km Stone, NH#2, Mathura-Delhi Road,
P.O. Chaumuhan, Mathura-281406 (U.P.)

Certificate

This is to certify that the above statements made by the candidate are
correct to the best of my/our knowledge and belief.

Supervisor
(Name of Supervisor)
Designation of Supervisor
Dept. of Computer Engg, & App.

Project Co-ordinator Program Co-ordinator

(Dr. Mayank Srivastava) (Mr. Jitesh Kumar Bhatia)
Associate Professor Assistant Professor
Dept. of Computer Engg, & App. Dept. of Computer Engg, & App.

Date:
 ACKNOWLEDGEMENT

I take this opportunity to express my sincere gratitude and respect to GLA University,

Mathura for providing me a platform to pursue my studies and carry out my final

year project. I have a great pleasure in expressing my deep sense of gratitude to Mr.

Vikash Sawan, Assistant Professor, GLA University ,Mathura for his constant

encouragement.

I consider it a privilege and honor to the express my sincere gratitude the to my guide

Mr. Vikash Sawan, Assistant professor , Department of Computer Science

and Engineering , for the valuable guidance throughout the tenure of this review.

I also extend my thanks to all the faculty of Computer Science and Engineering who

directly or indirectly encouraged me.

Finally, I would like to thank my parents and friends for all their moral support they

have given me during the completion of this work.

Sign Sign
Ashutosh Sarkar Navneet Singh
201500169 201500433

Sign Sign
Himanshu Gupta Amit Kumar Yadav
201500294 201500081
 ABSTRACT

This project presents a comprehensive exploration of machine learning methodologies

for stock price prediction, addressing the inherent complexities of financial markets.
Leveraging historical stock data alongside pertinent financial indicators, we construct
predictive models to forecast stock prices. Our approach encompasses a diverse array
of machine learning algorithms, including linear regression, decision trees, random
forests, and recurrent neural networks (RNNs), each offering unique insights into
market dynamics. Additionally, we integrate sentiment analysis of news articles and
social media data to capture market sentiment, augmenting prediction accuracy.

Through extensive experimentation with real-world stock market data, we

meticulously evaluate the performance of these machine learning techniques in terms
of predictive accuracy and scalability. Our findings underscore the efficacy of
machine learning in forecasting stock prices, demonstrating notable improvements
over traditional methods. Furthermore, we elucidate the significance of sentiment
analysis as a complementary tool for enhancing prediction accuracy, particularly in
capturing market sentiment nuances.

This project contributes substantively to the advancement of stock market prediction

methodologies by offering a comprehensive framework for leveraging machine
learning techniques. By providing valuable insights into market behavior and trends,
our approach empowers investors and financial analysts to make more informed
decisions. Ultimately, this research facilitates a deeper understanding of the intricate
interplay between data-driven analysis, market sentiment, and stock price dynamics,
thereby enriching the landscape of financial forecasting and investment strategies.
 CONTENTS
Declaration ii
Certificate ii
Acknowledge iii
Abstract iv
List of figures v
List of Tables vi

CHAPTER 1 Introduction 1
1.1 introduction 1
1.2 overview 2
1.3 motivation 2
1.4 objective 3
1.5 Issue and challenges 4
1.6 Organization of the project 4
1.7 Outcome of the project 4
CHAPTER 2 Software Requirement Analysis 5
2.1 Hardware requirements 5
2.2 Software specifications 5
2.3 Software requirements 6
2.3.1 Anaconda Distribution
2.3.2 Python Libraries
2.3.3 libraries used in the code and their details
CHAPTER 3 Software Design 7
3.1 Architectural Overview 7
3.1.1 Components 7
3.1.2 Interaction 8
3.1.3 Technologies Used 8
3.2 Data flow diagram 9
3.2.1 Level 0 DFD 9
 3.2.3 Level 1 DFD 10
3.2.3 Data flow path 10
3.2.4 Data source store 10
3.2.5 External Entities 10
3.2.6 Data Flow Annotations 10
3.2.7 Data Flow Controls 10
3.2.8 Data Flow Optimization 10
3.3 Component Design 11
3.3.1 Data Collection Module 11
3.3.2 Preprocessing Module 11
3.3.3 Model Training Module 11
3.3.4 Prediction Generation Module 12
3.3.5 User Interface Module 12
3.4 Model Architecture 13
3.4.1 Algorithm Selection 13
3.4.2 Input Features 13
3.4.3 Model Architecture 14
3.4.4 Output Formats 14
3.5 Error Handling 14
3.5.1 Error Detection 14
3.5.2 Error Handling Strategies 15
3.6 User Interface Design 15
3.6.1 Design Principles 15
3.6.2 Components and Features 15
3.6.3 Responsive Design 16
3.7 RNN-based Approach 16
3.7.1 Recurrent Neural Networks (RNNs) 16
3.7.2 Long Short-Term Memory Networks (LSTMs) 17
3.7.3 Application in Stock Price Prediction 17
3.7.4 Benefits and Challenges 17
3.7.5 Future Directions

3.8 Conclusion 18

CHAPTER 4 Implementation and User Interface 19

 4.1 Moving Averages (MA) 21
4.2 Long Short-Term Memory Networks (LSTM) 22
4.3 Algorithm 23
4.4 Here is the working of model 24
4.4.1 Data Preparation 24
4.4.2 Model Creation 24
4.4.3 Model Compilation 24
4.4.4 Model Training 25
4.4.5 Model Evaluation 26
4.4.6 Result Visualization 27
4.5 Web app 28
4.5.1 Code Snipits (fig 4.2) 28
4.5.2 Input Companies name (fig 4.3) 28
4.5.2 Output (fig 4.4) 29
CHAPTER 5 Software Testing 30
5.1 Importance of Software Testing 30
5.2 Testing Strategies 30
5.3 Testing Methodologies 30
5.4 Testing Techniques 31
5.5 Testing Tools and Frameworks 31
5.6 Testing Best Practices 32
5.7 Testing Metrics and Reporting 32
5.8 Challenges and Considerations 32
5.9 Future Directions in Software Testing 33
CHAPTER 6 Litreture review 34
CHAPTER 7 Conclusion 36
7.1 Summary of Findings 36
7.2 Achievement of Objectives 36
7.3 Contributions to Knowledge 36
7.4 Implications for Practice 36
7.5 Limitations and Future Research Directions 37
7.6 Conclusion 38
REFRENCES 39
List of Figures

3.1 Fig 3.1 - Architectural Design 7

3.2 Fig 3.2 - Level 0 DFD 9
3.3 Fig 3.3 – Level 1 DFD 9
4.1 Code Snipits (fig 4.2) 28
4.2 Input Companies name (fig 4.3) 28
4.4 Output (fig 4.4) 29
List of

1.1Basic writing technique rules 37

3.1Tables in illustrating data 32
 Chapter 1
INTRODUCTION
1.1 Introduction
Stock price prediction is a vital component of financial decision-making for investors,
businesses, and financial institutions alike. The ability to forecast stock prices
accurately enables stakeholders to anticipate market movements, identify investment
opportunities, and mitigate risks effectively. However, traditional methods of stock
price
prediction often face challenges in capturing the intricate dynamics and uncertainties
inherent in financial markets.

Conventional approaches, which typically rely on statistical methods or simple time

series models, may struggle to adequately capture the complex patterns and
dependencies present in stock price data. Factors such as market sentiment,
geopolitical events, and economic indicators contribute to the volatility and
unpredictability of stock prices, making accurate prediction a formidable task.
In recent years, advancements in deep learning techniques have shown promise in
enhancing the accuracy of stock price prediction. Specifically, Long Short-Term
Memory (LSTM) and Recurrent Neural Network (RNN) architectures have emerged
as powerful tools for modeling sequential data and capturing temporal dependencies.

LSTM and RNN models possess the ability to learn from historical data and capture
long-term dependencies, making them well-suited for time series forecasting tasks
such as stock price prediction. By leveraging the inherent memory capabilities of these
neural network architectures, researchers and practitioners aim to improve the
accuracy and robustness of stock price predictions.
In this context, this paper explores the application of LSTM and RNN models in the
domain of stock price prediction. We aim to investigate the potential of deep learning
techniques in overcoming the limitations of traditional methods and enhancing

Dept. of CEA, GLAU, 1

 Chapter Introducti
prediction accuracy in financial markets. Through empirical analysis and
experimentation, we seek to evaluate the effectiveness of LSTM and RNN models in
capturing the complexities of stock price data and providing accurate forecasts.

Overall, this study aims to contribute to the growing body of research on the
application of deep learning in financial forecasting, with a specific focus on stock
price prediction using LSTM and RNN architectures.

1.2 Overview
This project focuses on developing a robust stock price prediction system using
machine learning techniques. By leveraging historical stock data, financial indicators,
and sentiment analysis of news articles and social media data, we aim to forecast stock
prices accurately. The project employs various machine learning algorithms such as
linear regression, decision trees, random forests, and recurrent neural networks
(RNNs) to analyze market trends and patterns. Additionally, we propose a hybrid
approach that combines content-based filtering and collaborative filtering techniques
for enhanced prediction accuracy. Real-world stock market data is utilized for
comprehensive experimentation and evaluation of the proposed methods. The project
aims to contribute to the advancement of stock market prediction methodologies by
providing insights into market dynamics and improving prediction accuracy, thereby
empowering investors and financial analysts to make more informed decisions in the
dynamic realm of financial markets.

1.3 Motivation
The motivation behind this project stems from the critical importance of accurate
stock price prediction in the financial industry. Investors and financial analysts rely
heavily on predictive models to make informed decisions regarding buying, selling,
and holding stocks. However, traditional methods often fall short in capturing the
complexities and nuances of dynamic market behavior. By harnessing the power of
machine learning, we aim to revolutionize stock price prediction by leveraging vast
amounts of historical data and incorporating advanced algorithms to identify
underlying patterns and trends. Additionally, the integration of sentiment analysis
allows us to capture market

Dept. of CEA, GLAU, 2

 Chapter Introducti
sentiment, providing deeper insights into investor behavior and market dynamics.
Ultimately, our motivation lies in developing a reliable and scalable stock price
prediction system that empowers stakeholders with actionable insights, enabling them
to navigate the complexities of financial markets with confidence and precision.

1.4 Objective
The objective of this project is to develop a robust stock price prediction system using
machine learning techniques. Our primary aim is to leverage historical stock data and
relevant financial indicators to train predictive models capable of accurately
forecasting stock prices. To achieve this, we plan to explore various machine learning
algorithms, including linear regression, decision trees, random forests, and recurrent
neural networks (RNNs), to analyze market trends and patterns effectively.
Additionally, we intend to integrate sentiment analysis of news articles and social
media data to capture market sentiment and enhance prediction accuracy. Through
comprehensive experimentation with real-world stock market data, we aim to evaluate
the performance of these machine learning techniques rigorously. Ultimately, our
objective is to contribute to the advancement of stock market prediction
methodologies, providing valuable insights into market dynamics and improving
prediction accuracy. This will empower investors and financial analysts with
actionable insights to navigate the complexities of financial markets confidently.

1.5 Issues and Challenges

Developing a robust stock price prediction system using machine learning is fraught
with challenges. The dynamic and often unpredictable nature of financial markets
poses a significant hurdle, requiring sophisticated algorithms capable of capturing
complex market dynamics. Data preprocessing, feature selection, and model training
are further complicated by the vast amount of available data and the need to
incorporate diverse sources such as historical stock prices, financial indicators, and
sentiment analysis from news articles and social media. Overfitting and underfitting
must be carefully managed to ensure models generalize well to unseen data.
Moreover, the continuous evolution of market conditions necessitates ongoing model
refinement and adaptation. Interpreting prediction results and translating them into
actionable insights for investors and

Dept. of CEA, GLAU, 3

 Chapter Introducti
financial analysts also present significant challenges. Addressing these hurdles
demands a multidisciplinary approach and the utilization of advanced techniques to
develop a reliable and effective prediction system.

1.6 Organization of the Project Report

The report is organized as follows, Chapter 1 provides a brief introduction about the
RNN’s and proposed models for the stock price prediction . Chapter 2 describes the
system requirement or existing system related to this project.
Chapter 3 gives detailed brief about the system architecture and working of each
module of the system. Chapter 4 discusses the implementation details of the system
and the results of the work. Chapter 5 concludes the work by summarizing the results.

1.7 Outcome of Project

The outcome of this project is a robust and accurate stock price prediction system
leveraging machine learning techniques. Through extensive experimentation and
refinement, the developed system demonstrates the ability to forecast stock prices
with a high degree of accuracy, thereby providing valuable insights to investors and
financial analysts. By integrating historical stock data, financial indicators, and
sentiment analysis from news articles and social media, the system captures complex
market dynamics and incorporates them into its predictions. The system's performance
is evaluated rigorously using real-world stock market data, highlighting its
effectiveness in navigating the dynamic nature of financial markets.

Dept. of CEA, GLAU, 4

 Chapter 2
SYSTEM REQUIREMENT

2.1 Hardware Requirements

 A PC with Windows/Linux OS
 Processor with 1.7-2.4gHz speed
 Minimum of 8gb RAM
 2gb Graphic card

2.2 Software Specification

 Text Editor (jupyter)
 VS Code
 Python libraries

2.3 Software Requirements

2.3.1 Anaconda distribution:
Anaconda is a free and open-source distribution of the Python programming
languages for scientific computing (data science, machine learning applications, large-
scale data processing, predictive analytics, etc.), that aims to simplify
packagemanagement system and deployment. Package versions are managed by the
packagemanagement system conda. The anaconda distribution includes data-science
packagessuitable for Windows, Linux and MacOS.

2.3.2 Python libraries:

In this project, several Python libraries are utilized to facilitate the development of a
robust stock price prediction system. Firstly, Pandas is employed for efficient data
manipulation and analysis, particularly in handling

Dept. of CEA, GLAU, 5

 Chapter System
time series data such as historical stock prices and financial indicators. NumPy
complements Pandas by providing essential numerical computing capabilities
and efficient array operations. Scikit-learn is essential for machine learning tasks,
offering various algorithms for model training, evaluation, and hyperparameter tuning.
For deep learning models like recurrent neural networks (RNNs) or long short-term
memory networks (LSTMs), TensorFlow or PyTorch is utilized. Matplotlib or
Seaborn are employed for data visualization, enabling the creation of insightful plots
to visualize stock price trends and model performance. Additionally, NLTK or spaCy
is used for natural language processing tasks, such as sentiment analysis of news
articles and social media data. Lastly, libraries like Yahoo Finance API or Alpha
Vantage API enable access to historical stock price data, while Flask or Django
facilitate the deployment of prediction models through web applications or APIs.
These libraries collectively provide a comprehensive toolkit for building and
deploying stock price prediction systems in Python.

2.3.3 libraries used in the code and their details

NumPy:
 Purpose: NumPy is a fundamental package for scientific computing in
Python, providing support for large, multi-dimensional arrays and matrices,
along with a collection of mathematical functions to operate on these arrays.
 Use Cases: Numerical computations, linear algebra, Fourier transforms, and
random number generation.

Pandas:
 Purpose: Pandas is a powerful data manipulation and analysis library, built on
top of NumPy. It offers data structures like Series and DataFrame, along with
functions for data cleaning, reshaping, merging, and time series analysis.
 Use Cases: Data cleaning, data wrangling, exploratory data analysis, and time
series analysis.

Dept. of CEA, GLAU, 6

Chapter System
Matplotlib:
 Purpose: Matplotlib is a plotting library for creating static, interactive, and
animated visualizations in Python. It provides a MATLAB-like interface and
supports a wide range of plots, including line plots, scatter plots, bar plots,
histograms, and heatmaps.
 Use Cases: Data visualization, exploratory data analysis, and communicating
results.

Seaborn:
 Purpose: Seaborn is a statistical data visualization library based on Matplotlib,
offering a higher-level interface for creating attractive and informative
statistical graphics. It provides functions for visualizing distributions,
relationships, and categorical data.
 Use Cases: Statistical data visualization, exploring relationships between
variables, and creating publication-quality plots.

Scikit-learn:
 Purpose: Scikit-learn is a machine learning library that provides simple and
efficient tools for data mining and data analysis. It includes a wide range of
algorithms for classification, regression, clustering, dimensionality reduction,
and model selection.
 Use Cases: Machine learning tasks such as classification, regression,
clustering, and model evaluation.

TensorFlow:
 Purpose: TensorFlow is an open-source machine learning framework
developed by Google for building and training deep learning models. It
provides tools for implementing neural networks, including high-level APIs
like Keras and low-level operations for custom model development.
 Use Cases: Deep learning, neural networks, natural language processing,
computer vision, and reinforcement learning.

Dept. of CEA, GLAU, 7

 Chapter System
PyTorch:
 Purpose: PyTorch is a deep learning framework developed by Facebook's AI
Research lab, known for its dynamic computational graph construction. It
offers flexible and intuitive APIs for building and training neural networks,
with support for automatic differentiation.
 Use Cases: Deep learning, neural networks, natural language processing,
computer vision, and reinforcement learning.

Keras:
 Purpose: Keras is a high-level neural networks API written in Python, capable
of running on top of TensorFlow, Theano, or Microsoft Cognitive Toolkit
(CNTK). It provides a user-friendly interface for building and training deep
learning models with minimal code.
 Use Cases: Rapid prototyping, building and training deep learning models
with ease.

Dept. of CEA, GLAU, 8

 Chapter 3
SYSTEM DESIGN

3.1 Architectural Overview

The architectural overview provides a high-level description of the stock price
prediction system's architecture, outlining its major components and their interactions.
This section serves as a foundational guide for understanding the system's structure
and functionality.

Fig 3.1 - Architectural Design

3.1.1 Components
 Data Collection Module: Responsible for gathering historical stock price
data, financial indicators, and sentiment analysis data from various sources
such as financial databases, APIs, and web scraping.
 Pre-processing Module: Handles data cleaning, feature engineering, and
normalization tasks to prepare the raw data for model training and prediction.

Dept. of CEA, GLAU, 9

Chapter System
 Model Training Module: Trains machine learning models using the pre-

processed data to predict future stock prices. This module encompasses

algorithm selection, hyperparameter tuning, and model evaluation.
 Prediction Generation Module: Generates stock price predictions based on
the trained models and user-defined parameters, such as forecast horizon and
input features.
 User Interface Module: Provides a user-friendly interface for users to interact
with the system, visualize prediction results, customize parameters, and access
additional functionalities.

3.1.2 Interactions
 Data flows from the Data Collection Module to the Pre-processing Module,
where it undergoes cleaning and transformation before being passed to the
Model Training Module.
 Trained models from the Model Training Module are deployed within the
Prediction Generation Module, which generates predictions based on user
input and model outputs.
 The User Interface Module interacts with all other modules, allowing users to
input data, customize settings, view prediction results, and access system
functionalities.

3.1.3 Technologies Used

 Python: The primary programming language used for system development,
leveraging its extensive libraries for data processing, machine learning, and
web development.
 Pandas, NumPy: For data manipulation and numerical computing tasks,
facilitating efficient handling of large datasets and array operations.
 Scikit-learn, TensorFlow, PyTorch: For implementing machine learning
algorithms and deep learning models for stock price prediction tasks.
 Flask, Django: Web frameworks used for developing the user interface and
deploying the system as web applications or APIs.
 Python and streamlit : For designing and implementing the user interface,
providing a visually appealing and interactive experience for users.

Dept. of CEA, GLAU, 1

Chapter System
3.2 Data Flow Diagram
The data flow diagram (DFD) illustrates the flow of data within the stock price
prediction system, depicting how raw data is processed, transformed, and utilized
throughout the prediction pipeline. This diagram serves as a visual representation of
the system's data flow, highlighting the interactions between different components and
subsystems.

Fig 3.2 - Level 0 DFD

Fig 3.3 – Level 1 DFD

3.2.1 Level 0 DFD

 The Level 0 DFD expands on the context diagram, breaking down the system
into major components or subprocesses and illustrating the flow of data
between them.
3.2.2 Level 1 DFD
 Level 1 DFDs further decompose the subprocesses identified in the Level 0
DFD into more detailed components, illustrating the specific data
transformations and interactions within each subprocess.
3.2.3 Data Flow Paths
 Data flow paths trace the movement of data through the system, identifying
inputs, outputs, and intermediate data transformations at each stage of the
prediction pipeline.

Dept. of CEA, GLAU, 1

Chapter System
3.2.4 Data Stores
 Data stores represent repositories of data within the system, such as databases
or files, where raw data is stored, processed, and accessed by different
components.
3.2.5 External Entities
 External entities represent sources of input or destinations of output for the
system, including users, data sources, and external systems interfacing with the
prediction system.
3.2.6 Data Flow Annotations
 Data flow annotations provide additional context and information about the
data flowing through the system, including data types, formats, and processing
steps applied to the data.
3.2.7 Data Flow Controls
 Data flow controls represent mechanisms or procedures within the system that
regulate the flow of data, such as data validation, error handling, and access
controls.
3.2.8 Data Flow Optimization
 Data flow optimization strategies are discussed to streamline data processing
and improve system performance, including parallel processing, caching, and
data compression techniques.

3.3 Component Design

The component design section delves into the detailed design of each major
component of the stock price prediction system, outlining their functionalities,
interfaces, and interactions with other components. This section serves to provide a
comprehensive understanding of the system's architecture and internal workings.
3.3.1 Data Collection Module
 Functionality: The Data Collection Module is responsible for gathering
historical stock price data, financial indicators, and sentiment analysis data
from various sources such as financial databases, APIs, and web scraping.

Dept. of CEA, GLAU, 1

Chapter System
 Interfaces: It interfaces with external data sources and APIs to retrieve raw

data, which is then passed to the Preprocessing Module for further processing.
 Interactions: The Data Collection Module interacts with the Preprocessing
Module to provide raw data for cleaning and transformation.

3.3.2 Preprocessing Module

 Functionality: The Preprocessing Module handles data cleaning, feature
engineering, and normalization tasks to prepare the raw data for model training
and prediction.
 Interfaces: It receives raw data from the Data Collection Module and
interfaces with the Model Training Module to provide preprocessed data for
model training.
 Interactions: The Preprocessing Module interacts with the Data Collection
Module to receive raw data and with the Model Training Module to provide
preprocessed data for training.

3.3.3 Model Training Module

 Functionality: The Model Training Module trains machine learning models
using the preprocessed data to predict future stock prices. This module
encompasses algorithm selection, hyperparameter tuning, and model
evaluation.
 Interfaces: It receives preprocessed data from the Preprocessing Module and
interfaces with the Prediction Generation Module to deploy trained models for
generating predictions.
 Interactions: The Model Training Module interacts with the Preprocessing
Module to receive preprocessed data and with the Prediction Generation
Module to deploy trained models for prediction.

3.3.4 Prediction Generation Module

 Functionality: The Prediction Generation Module generates stock price
predictions based on the trained models and user-defined parameters, such as
forecast horizon and input features.

Dept. of CEA, GLAU, 1

Chapter System
 Interfaces: It receives input parameters from the User Interface Module and

interfaces with the Model Training Module to deploy trained models for
generating predictions.
 Interactions: The Prediction Generation Module interacts with the User
Interface Module to receive input parameters and with the Model Training
Module to deploy trained models for prediction.

3.3.5 User Interface Module

 Functionality: The User Interface Module provides a user-friendly interface
for users to interact with the system, visualize prediction results, customize
parameters, and access additional functionalities.
 Interfaces: It interfaces with all other modules to facilitate user input,
parameter customization, and visualization of prediction results.
 Interactions: The User Interface Module interacts with all other modules to
receive input parameters, display prediction results, and facilitate user
interaction with the system.
This component design section offers a detailed breakdown of each major component
of the stock price prediction system, elucidating their functionalities, interfaces, and
interactions within the overall system architecture.

3.4 Model Architecture

The model architecture section provides an in-depth explanation of the architecture of
machine learning models utilized for stock price prediction. This section covers
various aspects of model design, including algorithm selection, input features, model
architecture, and output formats.

3.4.1 Algorithm Selection

 Regression Algorithms: Evaluate and select appropriate regression algorithms
for stock price prediction, considering factors such as performance,
interpretability, and suitability for time series data.

Dept. of CEA, GLAU, 1

Chapter System
 Ensemble Methods: Explore ensemble methods such as random forests,

gradient boosting, and stacking to combine predictions from multiple base

models and improve prediction accuracy.
 Deep Learning Models: Consider deep learning models such as recurrent
neural networks (RNNs), long short-term memory networks (LSTMs), and
convolutional neural networks (CNNs) for capturing complex temporal
patterns in stock price data.

3.4.2 Input Features

 Historical Price Data: Include historical stock prices, volume, and trading
indicators such as moving averages, and exponential moving averages (EMA)
as input features.
 Financial Indicators: Incorporate relevant financial indicators such as price-
to-earnings ratio (P/E), earnings per share (EPS), and market capitalization to
capture fundamental aspects of stock performance.
 Sentiment Analysis: Integrate sentiment analysis features derived from news
articles, social media data, and investor sentiment indices to capture market
sentiment and its impact on stock prices.

3.4.3 Model Architecture

 Feature Engineering: Perform feature engineering techniques such as lag
features, rolling windows, and differencing to capture temporal dependencies
and trends in stock price data.
 Model Complexity: Experiment with different model architectures, including
simple linear models, tree-based models, and deep learning architectures, to
strike a balance between model complexity and prediction accuracy.
 Hyperparameter Tuning: Fine-tune model hyperparameters using techniques
such as grid search, random search, or Bayesian optimization to optimize
model performance and generalization.
3.4.4 Output Formats
 Point Estimates: Provide point estimates of future stock prices, indicating the
predicted value of the stock price at a specific time point.

Dept. of CEA, GLAU, 1

Chapter System
 Confidence Intervals: Calculate confidence intervals around point estimates

to quantify the uncertainty associated with each prediction, providing a

measure of prediction reliability and risk assessment.

3.5 Error Handling

Error handling is a critical aspect of the system design to ensure robustness, reliability,
and fault tolerance. This section discusses the strategies and mechanisms employed
for detecting, handling, and recovering from errors and exceptions that may occur
during system operation.

3.5.1 Error Detection

 Input Validation: Implement input validation mechanisms to ensure that user
inputs and external data sources are properly validated and sanitized to prevent
injection attacks, data corruption, or malformed data issues.
 Data Integrity Checks: Perform data integrity checks and validation routines
to verify the integrity and consistency of data, detecting errors such as data
corruption, missing values, or inconsistencies.
3.5.2 Error Handling Strategies
 Exception Handling: Implement exception handling mechanisms to catch and
handle runtime exceptions, errors, and failures gracefully, preventing system
crashes and ensuring uninterrupted operation.

3.6 User Interface Design

The user interface (UI) design is crucial for the stock price prediction system, as it
directly impacts user experience and usability. This section discusses the design
principles, components, and features of the user interface, ensuring it is intuitive,
informative, and efficient for users to interact with.

Dept. of CEA, GLAU, 1

Chapter System
3.6.1 Design Principles
 User-Centered Design: Follow a user-centered design approach, focusing on
the needs, preferences, and behaviors of users to create an interface that is easy
to use and understand.
 Consistency: Maintain consistency in design elements such as layout, color
scheme, typography, and navigation across all pages and components to
provide a cohesive and familiar user experience.
 Accessibility: Ensure the UI is accessible to users with disabilities,
incorporating features such as keyboard navigation, screen reader
compatibility, and high contrast modes.
 Simplicity: Keep the UI simple and uncluttered, avoiding unnecessary
elements or complexity that may confuse or overwhelm users.

3.6.2 Components and Features

 Interactive Charts: Provide interactive charts and graphs for visualizing
historical stock prices, prediction trends, and performance metrics, enabling
users to analyze data and make informed decisions.
 Customization Options: Allow users to customize prediction parameters,
input features, forecast horizons, and model selection to tailor predictions to
their specific needs and preferences.
 Historical Data Analysis: Enable users to explore historical stock price data,
perform trend analysis, and compare performance metrics to gain insights into
past market behavior.
 Prediction Visualization: Display predicted stock price trends, confidence
intervals, and model performance metrics in a clear and understandable format,
helping users interpret and act upon prediction results effectively.

3.6.3 Responsive Design

 Mobile Responsiveness: Ensure the UI is responsive and optimized for mobile
devices, providing a consistent and user-friendly experience across desktop,
tablet, and smartphone devices.

Dept. of CEA, GLAU, 1

Chapter System
 Cross-Browser Compatibility: Test the UI on different web browsers and
versions to ensure compatibility and consistent rendering.

3.7 RNN-based Approach

The RNN-based approach leverages recurrent neural networks (RNNs) and their
specialized variant, long short-term memory networks (LSTMs), to model sequential
data and capture temporal dependencies in stock price time series. This section
provides an overview of RNNs and LSTMs and their application in stock price
prediction.

3.7.1 Recurrent Neural Networks (RNNs)

Recurrent Neural Networks (RNNs) are a class of artificial neural networks designed
to process sequential data by maintaining an internal state or memory. Unlike
feedforward neural networks, RNNs have connections that form directed cycles,
allowing them to exhibit dynamic temporal behavior and capture dependencies across
time steps.
3.7.2 Long Short-Term Memory Networks (LSTMs)
Long Short-Term Memory Networks (LSTMs) are a specialized variant of RNNs
designed to address the vanishing gradient problem and capture long-term
dependencies in sequential data. LSTMs introduce gating mechanisms, including
input gates, forget gates, and output gates, to control the flow of information through
the network and selectively update or forget information over time.
3.7.3 Application in Stock Price Prediction
RNNs and LSTMs have been widely applied in stock price prediction tasks due to
their ability to model temporal dynamics and capture complex patterns in time series
data. By processing historical stock price data as sequential input sequences, RNNs
and LSTMs can learn to predict future price movements and identify trading
opportunities.
3.7.4 Benefits and Challenges
Benefits: RNNs and LSTMs can effectively capture temporal dependencies and
nonlinear relationships in stock price data, enabling more accurate and informative
predictions.

Dept. of CEA, GLAU, 1

 Chapter System
Challenges: However, training RNNs and LSTMs can be computationally intensive
and require careful tuning of hyperparameters to prevent overfitting and ensure model
stability.

3.7.5 Future Directions

As research in deep learning continues to advance, future directions for RNN-based
approaches in stock price prediction may include the exploration of attention
mechanisms, transformer architectures, and hybrid models combining RNNs with
other neural network architectures to further improve prediction accuracy and
robustness.
This section provides a brief overview of the RNN-based approach and its application
in stock price prediction, focusing on the architecture and capabilities of recurrent
neural networks (RNNs) and long short-term memory networks (LSTMs).

Here are some potential areas for future work in the context of stock price prediction
using deep learning:
 Data Augmentation: Explore data augmentation techniques to increase the
size of the training data set and improve the model's ability to generalize to
unseen data.
 Feature Engineering: Experiment with different feature engineering
techniques, such as adding technical indicators or incorporating external data
sources, to improve the model's performance.
 Model Architecture: Experiment with different model architectures, such as
adding more layers or using different types of layers, to improve the model's
ability to capture complex patterns in the data.
 Optimization: Experiment with different optimization techniques, such as
using different optimizers or adjusting the learning rate, to improve the
model's convergence and training time.
 Evaluation Metrics: Explore different evaluation metrics, such as using
different loss functions or incorporating additional metrics, to better measure
the model's performance.
 Real-time Prediction: Implement a real-time prediction system that can
continuously update the model's predictions as new data becomes available.

Dept. of CEA, GLAU, 1

Chapter System
 Risk Management: Incorporate risk management techniques, such as setting
stop-loss orders or adjusting position sizes, to minimize the risk of losses due
to market volatility.
 Robustness Testing: Perform robustness testing to evaluate the model's
performance under different market conditions, such as during periods of high
volatility or low liquidity.
 Explainability: Explore techniques to improve the model's explainability,
such as using attention mechanisms or visualizing the model's internal
representations.
 Scalability: Implement a scalable architecture that can handle large volumes
of data and provide fast predictions, making it suitable for real-world
applications.

These are just a few potential areas for future work in the context of stock price
prediction using deep learning. By exploring these areas, researchers and practitioners
can continue to improve the accuracy and reliability of stock price predictions,
ultimately helping investors make more informed decisions.

3.8 Conclusion
In conclusion, the stock price prediction system offers a powerful tool for investors,
analysts, and financial professionals to make informed decisions and gain insights into
future market trends. This chapter explored various aspects of system design,
including architectural overview, data flow, component design, model architecture,
error handling, scalability, and user interface design. By adopting a comprehensive
approach and prioritizing user-centric design principles, the system aims to deliver a
reliable, efficient, and user-friendly solution for predicting stock prices and analyzing
market trends, providing valuable insights and enhanced decision-making capabilities
across the financial industry.

Dept. of CEA, GLAU, 2

 Chapter 4
IMPLEMENTATION AND USER
INTERFACE

The Proposed System Make Use Different Algorithms and Methods for the
Implementation of RNN and Approach of Moving averages.

4.1 Moving Averages (MA)

Moving averages are a commonly used statistical technique in time series analysis and forecasting,
including in stock price prediction. The basic idea behind moving averages is to smooth out
fluctuations in data by calculating the average value of a subset of data points over a specified time
window, known as the "moving" window.
 Simple Moving Average (SMA): This is the most basic form of moving
average, where the average is calculated by taking the arithmetic mean of a
fixed number of data points within the window. SMA assigns equal weight to
all data points within the window.

 Exponential Moving Average (EMA): EMA gives more weight to recent data
points while still considering older data points. It exponentially decreases the
weight of older observations, making it more responsive to recent changes in
the data series.

Dept. of CEA, GLAU, 2

Chapter 4 Implementation and user interface

4.2 Long Short-Term Memory Networks (LSTM)

Long Short-Term Memory (LSTM) networks are a type of recurrent neural network
(RNN) architecture designed to overcome the limitations of traditional RNNs, such as
the difficulty of capturing long-term dependencies in sequential data. LSTMs are
particularly effective for modeling time series data like stock prices due to their ability
to maintain a memory state over time.
 Memory Cells: LSTMs contain special memory cells that can maintain
information over long sequences of data. These cells have three gates – input,
output, and forget gates – that regulate the flow of information, allowing the
network to selectively update or retain information over time.
 Long-Term Dependencies: LSTMs are capable of learning long-term
dependencies in sequential data by selectively storing and accessing relevant
information in the memory cells over multiple time steps.
 Training: LSTMs are trained using gradient-based optimization algorithms
such as backpropagation through time (BPTT), where the network learns to
adjust its parameters to minimize the difference between predicted and actual
values over a training dataset.

Fig 4.1 - LSTM(Long short term memory)

Dept. of CEA, GLAU, 2

Chapter Implementation and user
4.3 Algorithm
Input: Stock Name
Output: Predicted Price Graph of that Stock
1. Data Collection: Gather historical stock price data from various sources such
as financial databases, APIs, or web scraping techniques.
2. Data Preprocessing: Clean the raw data by handling missing values, outliers,
and inconsistencies. Perform feature engineering to extract relevant features
such as moving averages, technical indicators, and sentiment analysis scores.
3. Model Selection: Choose appropriate machine learning algorithms for stock
price prediction, such as regression models (e.g., linear regression, polynomial
regression), time series models (e.g., ARIMA, SARIMA), or deep learning
models (e.g., Long Short-Term Memory networks - LSTM).
4. Model Training: Split the data into training and testing sets. Train the selected
model on the training data, adjusting model parameters through techniques
such as cross-validation or hyperparameter tuning.
5. Model Evaluation: Evaluate the trained model's performance on the testing
data using appropriate evaluation metrics such as Mean Absolute Error
(MAE), Mean Squared Error (MSE), or Root Mean Squared Error (RMSE).
6. Prediction Generation: Use the trained model to make predictions on unseen or
future data, generating forecasts for future stock prices.
7. Post-Processing: Post-process the model predictions as needed, such as
converting them into actionable trading signals or visualizing them for
interpretation by users.
8. Feedback Loop: Continuously monitor the performance of the prediction
model and incorporate feedback to refine and improve the algorithm over time.

4.4 Here is the working of model

1. Data Preparation:
 Scale the training data set using a MinMaxScaler.
 Create input-output pairs for training by selecting sequences of 100 days of
scaled data.

Dept. of CEA, GLAU, 2

Chapter Implementation and user
2. Model Creation:
 Import necessary libraries like Dense, Dropout, LSTM, and Sequential from
Keras.
 Define a sequential model with LSTM layers, dropout layers, and a dense
output layer.
3. Model Compilation:
 Compile the model using the Adam optimizer and mean squared error loss
function.
4. Model Training:
 Fit the model on the input-output pairs with 50 epochs and a batch size of 32.
5. Model Evaluation:
 Print a summary of the model to review the architecture and parameters.
 Concatenate the last 100 days of training data with the testing data set.
 Scale the testing data set using the same scaler used for training.
 Create Print a summary of the model to review the architecture and parameters.
 input-output pairs for testing data by selecting sequences of 100 days.
 Predict the stock prices for the testing data set using the trained model.
6. Results Visualization:
 Transform the predicted and actual values back to their original scale.
 Plot a chart to visualize the predicted and actual stock prices over time.
This workflow outlines the steps involved in training a deep learning model for stock
price prediction, from data preparation and model creation to evaluation, visualization,
and model saving.
4.4.1 Data Preparation
 Data collection and preprocessing are essential steps in stock price prediction.
 Data sources for stock price prediction include historical stock prices, financial
reports, news articles, and social media posts.
 Preprocessing involves cleaning, normalizing, and transforming the data to
make it suitable for ML models.

Dept. of CEA, GLAU, 2

Chapter Implementation and user
 Scaling the training data set using a MinMaxScaler is necessary to ensure that
the ML model can learn the relationships between the features and the target
variable.
 Creating input-output pairs for training involves selecting a sequence of
historical stock prices as input and the corresponding stock price as output.
 The input-output pairs are used to train the ML model to predict the output
based on the input.
 The quality of data preparation significantly impacts the accuracy and
reliability of the ML model's predictions.

4.4.2 Model Creation

 Importing necessary libraries from Keras is the first step in creating a deep
learning model for stock price prediction.
 Defining a sequential model with LSTM layers, dropout layers, and a dense
output layer is essential for capturing temporal dependencies and generalizing
patterns in the data.
 LSTM (Long Short-Term Memory) layers are well-suited for sequential data
like stock prices, as they can retain information over long sequences.
 Dropout layers help prevent overfitting by randomly setting a fraction of input
units to zero during training.
 A dense output layer is used to produce the final prediction based on the
learned patterns in the data.

4.4.3 Model Compilation

 Compiling the model is a crucial step in configuring the training process for
the deep learning model.
 Using the Adam optimizer is a popular choice for optimizing deep learning
models due to its adaptive learning rate and momentum.
 The mean squared error loss function is commonly used for regression tasks
like stock price prediction, as it measures the average squared difference
between the predicted and actual values.

Dept. of CEA, GLAU, 2

Chapter Implementation and user
 Configuring the model with the Adam optimizer and mean squared error loss
function sets the stage for efficient training and accurate predictions.

4.4.4 Model Training

 Fitting the model on the input-output pairs is the process of training the deep
learning model on the training data.
 Training the model involves optimizing the model parameters to minimize the
difference between the predicted and actual values.
 Using 50 epochs means the model will iterate over the entire training dataset
50 times during training.
 A batch size of 32 indicates that the model will update its parameters after
processing 32 samples from the training data.
 Training the model with multiple epochs and a suitable batch size helps
improve the model's ability to learn complex patterns in the data.

4.4.5 Model Evaluation

 Printing a summary of the model provides an overview of the model
architecture, including the number of layers, the number of parameters, and the
input and output shapes.
 Concatenating the last 100 days of training data with the testing data set allows
for evaluating the model's performance on unseen data.
 Scaling the testing data set using the same scaler used for training ensures that
the testing data is in the same scale as the training data.
 Creating input-output pairs for testing data by selecting sequences of 100 days
enables evaluating the model's ability to predict stock prices based on
historical data.
 Predicting the stock prices for the testing data set using the trained model
allows for evaluating the model's performance on unseen data.

Dept. of CEA, GLAU, 2

Chapter Implementation and user
4.4.6 Result Visualization
 Transforming the predicted and actual values back to their original scale is
essential for interpreting the model's predictions in the context of actual stock
prices.
 Plotting a chart to visualize the predicted and actual stock prices over time
provides a clear visual comparison of the model's forecasts against the real
data.
 The chart helps in assessing the model's performance, identifying trends, and
understanding how well the model captures the underlying patterns in the
stock prices.
 Visualizing the predicted and actual stock prices allows for easy interpretation
and comparison, aiding in decision-making based on the model's predictions.

Here is a diagram of working model (fig 4.1)

Dept. of CEA, GLAU, 2

Chapter Implementation and user
4.5 Web app
4.5.1 Code Snipits (fig 4.2)

4.5.1 Input Companies name (fig 4.3)

Dept. of CEA, GLAU, 2

Chapter Implementation and user
4.5.2 Output (fig 4.4)

Dept. of CEA, GLAU, 2

 Chapter 5
SOFTWARE TESTING
Software testing focuses on the critical aspect of software testing in the development
lifecycle of the stock price prediction system. This chapter discusses various testing
strategies, methodologies, and techniques employed to ensure the reliability,
functionality, and performance of the system.

5.1 Importance of Software Testing

 Reliability: Testing ensures that the system functions reliably under different
conditions and scenarios, minimizing the risk of errors, bugs, and failures.
 Quality Assurance: Testing verifies that the system meets specified
requirements and adheres to quality standards, enhancing user satisfaction and
confidence in the system.
 Risk Mitigation: Testing identifies potential issues and vulnerabilities early in
the development process, allowing for timely resolution and risk mitigation.
 Compliance: Testing ensures that the system complies with regulatory
requirements, industry standards, and best practices, reducing legal and
financial risks.

5.2 Testing Strategies

 Unit Testing: Test individual components or modules of the system in
isolation to validate their correctness and functionality.
 Integration Testing: Verify the interactions and interfaces between different
system components to ensure seamless integration and interoperability.
 System Testing: Validate the system as a whole against its specified
requirements and use cases to ensure overall functionality and performance.
 Regression Testing: Repeatedly test the system after modifications or updates
to detect and prevent regression bugs and unintended side effects.

Dept. of CEA, GLAU, 3

Chapter Software
 Acceptance Testing: Obtain user feedback and validation to ensure that the

system meets user expectations and requirements.

5.3 Testing Methodologies

 Black Box Testing: Test the system based on its external behavior and inputs,
without knowledge of its internal implementation details.
 White Box Testing: Test the system based on its internal structure and logic,
with access to source code and implementation details.
 Gray Box Testing: Combine elements of black box and white box testing,
leveraging both external behavior and internal knowledge to design test cases.

5.4 Testing Techniques

 Functional Testing: Verify that the system functions correctly according to its
specified requirements and functional specifications.
 Non-Functional Testing: Evaluate aspects such as performance, scalability,
security, and usability to ensure the system meets non-functional requirements.
 Automated Testing: Use automated testing tools and frameworks to
streamline the testing process, increase test coverage, and accelerate feedback
loops.
 Manual Testing: Conduct manual testing by human testers to simulate real-
world user interactions and validate system behavior in diverse scenarios.

5.5 Testing Tools and Frameworks

 Unit Testing Frameworks: Utilize frameworks such as JUnit, pytest, or
NUnit for writing and executing unit tests for individual components.
 Automation Tools: Employ tools like Selenium, Appium, or Cypress for
automated functional testing of web applications and user interfaces.
 Performance Testing Tools: Use tools such as JMeter, LoadRunner, or
Gatling to measure and analyze system performance under different load
conditions.
 Security Testing Tools: Leverage tools like OWASP ZAP, Burp Suite, or
Nessus for identifying and mitigating security vulnerabilities in the system.

Dept. of CEA, GLAU, 3

Chapter Software
5.6 Testing Best Practices
 Test Planning and Documentation: Develop comprehensive test plans, test
cases, and test scripts to guide the testing process and ensure thorough
coverage.
 Continuous Integration and Continuous Testing: Integrate testing into the
development pipeline with continuous integration (CI) and continuous testing
(CT) practices to detect and address issues early.
 Bug Tracking and Management: Use bug tracking systems such as JIRA,
Bugzilla, or GitHub Issues to report, track, and prioritize identified issues
throughout the testing process.
 Collaboration and Communication: Foster collaboration and communication
between developers, testers, and stakeholders to facilitate knowledge sharing
and problem resolution.

5.7 Testing Metrics and Reporting

 Test Coverage: Measure test coverage metrics, such as statement coverage,
branch coverage, and path coverage, to assess the effectiveness and
completeness of testing.
 Defect Density: Calculate defect density metrics, such as defects per line of
code or defects per test case, to quantify the quality and reliability of the
system.
 Testing Progress: Track testing progress and status using metrics such as test
execution progress, test pass/fail rates, and defect resolution rates.
 Testing Reports: Generate testing reports and summaries to communicate test
results, findings, and recommendations to stakeholders and decision-makers.

5.8 Challenges and Considerations

 Resource Constraints: Address challenges related to limited time, budget,
and resources available for testing, prioritizing testing efforts based on risk and
criticality.
 Test Environment Management: Manage test environments effectively to
ensure consistency, reproducibility, and isolation of test scenarios from
production environments.

Dept. of CEA, GLAU, 3

Chapter Software
 Test Data Management: Secure and manage test data sets and environments

to ensure data privacy, confidentiality, and compliance with regulations.

5.9 Future Directions in Software Testing

 AI and Machine Learning: Explore the application of artificial intelligence
(AI) and machine learning (ML) techniques for automated test generation,
anomaly detection, and predictive analytics in software testing.
 DevOps and Testing Automation: Embrace DevOps principles and practices
to integrate testing seamlessly into the software development lifecycle (SDLC)
and automate testing processes for faster feedback and delivery.
 Shift-Left Testing: Shift testing activities earlier in the SDLC to detect and
address issues sooner, reducing the cost and impact of defects discovered later
in the development process.

Dept. of CEA, GLAU, 3

 Chapter 6
LITRETURE REVIEW
[1] Research on Stock Price Prediction Method Based on Convolutional Neural
Network, IEEE 2019- Sayavong Lounnapha et al. This paper intends for a prediction
model for stock price which is centered at the convolutional neural networks, that has
exceptional capability of learning on its own. The data set is taught and tested relating
the behaviours of both Convolutional Neural Networks and Thai stock market The
result shows that the model on grounds of Convolutional Neural Networks can
effectually recognize the altering trend in stock market price and envisage it which
provides significant allusion for stock price forecast. The accuracy of the prediction is
found to be elevated, and it could also be promoted in the field of finance.

[2] Enhancing Profit by Predicting Stock Prices using Deep Neural Networks, IEEE
2019-Soheila Abrishami, et al., The prediction of economic time series is quite a
herculean task, which has fascinated the attentiveness of many scholars and is
extremely vital for investors. This paper focuses on presenting a deep learning system,
which makes use of a range of facts for a part of the stocks on the NASDAQ exchange
to predict the value of the stock. This model has been trained on the smallest of data
for a particular stock and accurately estimates the concluding value of that stock for
multi- stepahead. It consists of an auto encoder in order to remove noise and makes
use of time series data engineering to syndicate the advanced features with the original
features. These new features are given to a Stacked LSTM Autoencoder for multistep-
ahead estimation of the stock concluding value. Further, this estimation is used by a
profit maximization approach to offer assistance on the right time for buying and
selling a particular stock. The results indicate that the suggested framework outclasses
the state of the art time series forecasting methodologies with respect to analytical
accuracy and effectiveness.

Dept. of CEA, GLAU, 3

Chapter 6 Litreture review

[3] An LSTM-Method for Bit-coin Price Prediction: A Case Study Yahoo Finance
Stock Market, IEEE 2019- Ferdiansyah et al., Bit-coin is a type of Cryptocurrency and
currently is one of a kind of investment on the stock market. Stock markets are
inclined by several risks. And bit-coin is one kind of crypto currency that keeps rising
in recent years, and sometimes suddenly falls without knowing influence on the stock
market. There’s a need for automation tools to predict bit-coin on the stock market
because of its fluctuations. This research study studies how to create mode prediction
bit-coin stock market prediction using LSTM. Before confirming the results the paper
tries to measure the results using RMSE (the Root Mean Square Error).The RMSE
will at all times be larger or equal to the MAE. The RMSE metric assesses how well a
model can calculate a continuous value. The method that is applied on this research to
predict Bit-coin on the stock market Yahoo finance can forecast the result above
$12600 USD for the next couple of days after prediction.

[4] Share Price Prediction using Machine Learning Technique, IEEE 2019-Jeevan B
et al., Lately stock market has been the talk of the town with more and more people
from academics and business showing interest in it. This paper mostly deals with the
approach towards predicting stock prices using RNN (Recurrent Neural Network) and
LSTM (Long Short Term Memory) on National Stock Exchange using numerous
elements such as the present-day market price as well as anonymous events. A
recommendation system along with models constructed on RNN and LSTM methods
are used in selecting the company is also mentioned in this paper.

[5] Stock Market Prediction Using Machine Learning Techniques, IEEE 2020-
Naadun Sirimevan et al., The Stock Market Prices play a crucial role in today’
economy. Researchers have discovered that social media platforms such as twitter and
web news tend to influence the decisionmaking process of any individual. In this
research behavioural reflex towards web news is taken into count to reduce the gap
and make the prediction much more accurate. Precise predictions were made for a
day, a week and two weeks here after.

Dept. of CEA, GLAU, 3

 Chapter 7
CONCLUSION
7.1 Summary of Findings
In this chapter, we summarize the key findings and insights gained from the
development and evaluation of the stock price prediction system. We revisit the
objectives of the project and highlight the main outcomes achieved through the
implementation and testing phases.

7.2 Achievement of Objectives

We reflect on the extent to which the objectives set forth at the outset of the project
have been met. This includes a discussion on the accuracy of predictions, the usability
of the system, and its scalability to handle large volumes of data and user requests.

7.3 Contributions to Knowledge

We examine the contributions of the project to the existing body of knowledge in the
field of stock price prediction. This includes advancements in machine learning
techniques, novel approaches to feature engineering, and insights into the dynamics of
financial markets.

7.4 Implications for Practice

We discuss the practical implications of the stock price prediction system for
stakeholders in the financial industry. This includes potential applications for
investors, financial analysts, and portfolio managers, as well as the integration of the
system into existing trading platforms and decision-making processes.

Dept. of CEA, GLAU, 3

Chapter 7 Conclusion

7.5 Limitations and Future Research Directions

We acknowledge the limitations of the stock price prediction system and identify
areas for future research and development. This includes addressing challenges related
to data quality, model interpretability, and the
incorporation of additional sources of information such as alternative data and market
sentiment.

7.6 Conclusion
In conclusion, we reflect on the journey of developing the stock price prediction
system and its significance in the context of financial markets. We highlight the
achievements, contributions, and implications of the project, and express our vision
for future advancements in the field of predictive analytics and algorithmic trading.

Dept. of CEA, GLAU, 3

 REFRENCES
[1] Naadun Sirimevan; I.G. U. H. Mamalgaha; Chandira Jayasekara; Y. S. Mayuran;
Chandimal Jayawardena (2020). Stock Market Prediction Using Machine Learning
Techniques in Faculty of Computing, Sri Lanka Institute of Information Technology,
Malabe, Sri Lanka.
DOI: 10.1109/ICAC49085.2019.9103381

[2] Sukhman Singh; Tarun Kumar Madan; Jitendra Kumar; Ashutosh Kumar Singh
(2020) Stock Market Forecasting using Machine Learning: Today and Tomorrow in
National Institute of Technology, Kurukshetra, Haryana, India.
DOI : 10.1109/ICICICT46008.2019.8993160

[3] Ferdiansyah Ferdiansyah; Siti Hajar Othman; Raja Zahilah Raja Md Radzi; Deris
Stiawan; Yoppy Sazaki; Usman Ependi(2020). A LSTM-Method for Bitcoin Price
Prediction: A Case Study Yahoo Finance Stock Market in School of Computing,
Universiti Teknologi Malaysia, Johor Bahru, Johor, Malaysia.
DOI : 10.1109/ICECOS47637.2019.8984499

[4] Meghna Misra; Ajay Prakash Yadav; Harkiran Kaur (2020). Stock Market
Prediction using Machine Learning Algorithms: A Classification Study in Department
of Computer Science and Engineering, (Deemed to be University, Patiala).
DOI : 10.1109/ICRIEECE44171.2018.9009178

[5] Mojtaba Nabipour , Pooyan Nayyeri , Hamed Jabani , Shahab S., (Senior Member,
Ieee), and Amir Mosavi (2020). Predicting Stock Market Trends Using Machine
Learning and Deep Learning Algorithms Via Continuous and Binary Data; a
Comparative Analysis.
DOI : 10.1109/ACCESS.2020.3015966

Dept. of CEA, GLAU, 3

 Referen
[6] B Jeevan, E Naresh, B P Vijaya kumar, Prashanth Kambli (2019). Share Price
Prediction using Machine Learning Technique in Department of Information Science
and Engineering, RIT, Bangalore 560 054.
DOI : 10.1109/CIMCA.2018.8739647

[7] Sumeet Sarode; Harsha G. Tolani; Prateek Kak; C S Lifna (2019). Stock Price
Prediction Using Machine Learning Techniques in Vivekanand Education Society's
Institute of
Technology, Mumbai, India. DOI: 10.1109/ISS1.2019.8907958

[8] Shao En Gao; Bo Sheng Lin; Chuin-Mu Wang(2019). Share Price Trend Prediction Using
CRNN with LSTM Structure in Department of Computer Science and Information
Engineering, National Chin-Yi University of Technology, Taichung, Taiwan.
DOI: 10.1109/IS3C.2018.00012

[9] Ishita Parmar; Navanshu Agarwal; Sheirsh Saxena; Ridam Arora; Shikhin Gupta;
Himanshu Dhiman; Lokesh Chouhan (2019). Stock Market Prediction Using Machine
Learning in Department of Computer Science and Engineering, National Institute Of
Technology, Hamirpur, INDIA.
DOI : 10.1109/ICSCCC.2018.8703332

[10] A.J.P. Samarawickrama; T.G.I. Fernando(2018). A recurrent neural network approach in

predicting daily stock prices an application to the Sri Lankan stock market in Department of
Computer Science, Faculty of Applied Sciences University of Sri Jayewardenepura,
Nugegoda, Sri Lanka.
DOI : 10.1109/ICIINFS.2017.8300345

[11] Chawalit Jeenanunta; Rujira Chaysiri; Laksmey Thong(2018). Stock Price Prediction
With Long Short-Term Memory Recurrent Neural Network in School of Management
Technology, Thammasat University, Thailand.
DOI : 10.1109/ICESIT-ICICTES.2018.8442069

[12] Mehak Usmani; Syed Hasan Adil; Kamran Raza; Syed Saad Azhar Ali (2016). Stock
market prediction using machine learning techniques in Iqra University, Karachi, PK.

Dept. of CEA, GLAU, 3

 Referen
DOI : 10.1109/ICCOINS.2016.7783235

[13] Jae Won Lee (2002). Stock price prediction using reinforcement learning in School of
Computer Science and Engineering, Sungshin Women's University, Seoul, South Korea.
DOI: 10.1109/ISIE.2001.931880

Dept. of CEA, GLAU, 4