A FINAL YEAR PROJECT REPORT ON

“STOCK PRICE PREDICTION USING MACHINE LEARNING MODEL”

SUBMITTED TO THE SAVITRIBAI PHULE PUNE UNIVERSITY, PUNE

IN THE PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE AWARD OF THE DEGREE

OF

BACHELOR OF ENGINEERING (COMPUTER

ENGINEERING)

SUBMITTED BY

ARYAN ASATI 72299812M

ADITYA MORE 72244658G
UMESH CHIMANE 72244655B
SAHIL BABAR 72244662E

DEPARTMENT OF COMPUTER ENGINEERING

PARVATIBAI GENBA MOZE COLLEGE OF ENGINEERING

BAIF ROAD, WAGHOLI, PUNE 412207

SAVITRIBAI PHULE PUNE UNIVERSITY

1
 2023 -2024
SAVITRIBAI PHULE PUNE UNIVERSITY
2024 -2025

CERTIFICATE

This is to certify that the project report entitles

“STOCK PRICE PREDICTION USING MACHINE LEARNING MODEL”

Submitted by

ARYAN ASATI 72299812M

ADITYA MORE 72244658G
UMESH CHIMANE 72244655B
SAHIL BABAR 72244662E

is a bonafide student of this institute and the work has been carried out by him/her under the
supervision of Prof. Jaishri Panchal and it is approved for the partial fulfillment of the
requirement of Savitribai Phule Pune University, for the award of the degree of Bachelor of
Engineering (Computer Engineering).

Prof. Jaishri Panchal Prof. S. D. Dhamdhere

Guide Head
Department of Computer Engineering Department of Computer Engineering

(Dr. N. S. Narwade)
Principal

2
 Place : Pune
Date :

ACKNOWLEDGEMENT

It is my pleasure to express a deep sense of gratitude to all those who helped me in the
completion of my final year project. I am highly indebted to Prof. Jaishri Panchal from
PGMCOE for her guidance, constant supervision, and for providing the necessary information
and support throughout the project.

I would also like to extend my gratitude to Prof. Shrikant Dhamdhere (Head of Computer
Department) for his cooperation and encouragement, which played a crucial role in providing the
required facilities for the project.

Lastly, I wish to thank and appreciate all my teachers and friends for their constructive
comments, suggestions, and guidance, as well as everyone who directly or indirectly helped me
complete this project successfully.

ARYAN ASATI

ADITYA MORE

UMESH CHIMANE

SAHIL BABAR

3
 ABSTRACT

The Stock Market, as we know, is volatile in nature and the prediction of the same is a
cumbersome task. Stock prices depend upon not only economic factors, but they relate to various
physical, psychological, rational and other important parameters. In this research work, the stock
prices are predicted using the Auto Regressive Integrated Moving Average (ARIMA) Model.
Stock price predictive models have been developed and run-on published stock data acquired
from Yahoo Finance. The experimental results lead to the conclusion that ARIMA Model can be
used to predict stock prices for a short period of time with reasonable accuracy.

Key Words: Machine Learning; Stock Market; Predictive Analysis; Financial Time Series
Forecasting

4
 TABLE OF CONTENTS

LIST OF ABBREVATIONS i
LIST OF FIGURES ii
LIST OF TABLES iii

INDEX

Sr.No. Title of Chapter Page

No.
01 Introduction
1.1 Motivation
1.2 Problem Definition
02 Literature Survey
03 Software Requirements Specification
3.1 Introduction
3.1.1 Project Scope
3.1.2 User Classes and Characteristics
3.1.3 Assumptions and Dependencies
3.2 Functional Requirements
3.2.1 System Feature 1(Functional Requirement)
3.2.2 System Feature2 (Functional Requirement)
……
……

5
 3.3 External Interface Requirements (If Any)
3.3.1 User Interfaces
3.3.2 Hardware Interfaces
3.3.3 Software Interfaces
3.3.4 Communication Interfaces
3.4 Non-functional Requirements
3.4.1 Performance Requirements
3.4.2 Safety Requirements
3.4.3 Security Requirements
3.4.4 Software Quality Attributes
3.5 System Requirements
3.3.1 Database Requirements
3.3.2 Software Requirements(Platform Choice)
5.3.3 Hardware Requirements
3.6 Analysis Models: SDLC Model to be applied
3.7 System Implementation Plan
04 System Design
4.1 System Architecture
4.2 Data Flow Diagrams
4.3 Entity Relationship Diagrams
4.4 UML Diagrams
05 Other Specification
5.1 Advantages
5.2 Limitations
5.3 Applications
06 Conclusions & Future Work
Appendix A: Problem statement feasibility assessment using, satisfiability analysis
and NP Hard, NP Complete or P type using modern algebra and relevant mathematical

models.

Appendix B: Details of the papers referred in IEEE format (given earlier) Summary
of the above paper in not more than 3-4 lines. Here you should write the seed idea of the

6
papers you had referred for preparation of this project report in the following format.
Example: Thomas Noltey, Hans Hanssony, Lucia Lo Belloz,”Communication Buses
for Automotive Applications” In Proceedings of the 3rd Information Survivability
Workshop (ISW-2007), Boston, Massachusetts, USA, October 2007. IEEE Computer
Society.

Appendix C: Plagiarism Report

References

7
 LIST OF ABBREVATIONS

Abbreviation Illustration
ARIMA Autoregressive Integrated Moving Average
API Application Programming Interface
LSTM Long Short-Term Memory (Neural Network)
API Application Programming Interface
RMSE Root Mean Square Error
ROI Return on Investment
BSE BOMBAY STOCK EXCHANGE
NSY National Stock Exchange

8
 LIST OF FIGURES

FIGURE ILLUSTRATION PAGE

NO

4.1 Architecture Diagram 18

4.2.1 Level-0 DFD 19

4.2.2 Level-1 DFD 20

4.4.1 Use-Case Diagram 22

4.4.2 Class Diagram 23

4.4.3 Activity Diagram 24

4.4.4 Sequence Diagram 25

LIST OF TABLES

2 Literature Survey 4

3.1.2 User Class & Characteristics 6

3.5.3 Hardware requirements 12

3.7 System implementation Plan 16

9
10
 CHAPTER 1
INTRODUCTION

One of the vital elements of a market economy is stock market. The reason behind this is mainly
because of the foundation it lays for public listed companies to gain capital via investors, who
invest to buy equity in the company. With the aid of refinements in the industries, stock market
is expanding rapidly. In order for the investors to gain returns (profits), they should take in
consideration the disparities involved in the stock market on regular basis. The stock market is
volatile in nature and the prediction of the same is not an easy task. Stock prices depend upon a
variety of factors including economic, physical, psychological, rational and other important
aspects. Although, the stock trend is difficult to predict, investors seem to find new techniques in
order to minimise the risk of investment and increase the probability of profiting from the
investments. The variability in stock market makes it an interesting field for researchers to forge
new forecasting models. Time-series analysis is an important subset of prediction algorithms and
functions. It is regarded as an apt tool for predicting the trends in stock market and logistics.
Before making any investment, an investor gathers intel on the past stock trends, periodic
changes and various other factors that affect the capital of a company. An ARIMA model is a
vibrant univariate forecasting method to project the future values of a time-series. Since, it is
essential to identify a model to analyse trends of stock prices with adequate information for
decision making, it is proposed to use the ARIMA model for stock price prediction.

Stock price prediction using the Autoregressive Integrated Moving Average (ARIMA) model is a
common technique in time series forecasting. ARIMA is a statistical method that models the
relationship between a series of observations and lags of that series.
Prediction will continue to be an interesting area of research making
researchers in the domain field always desiring to improve existing predictive models. The
reason is that institutions and individuals are empowered to make investment decisions and
ability to plan and develop effective strategy about their daily and future endeavors. Stock price
prediction is regarded as one of most difficult task to accomplish in financial forecasting due to
complex nature of stock market. The desire of many investors is to lay hold of any forecasting

11
method that could guarantee easy profiting and minimize investment risk from the stock market.
This remains a motivating factor for researchers to evolve and develop new predictive models

1.1 MOTIVATION

 The motivation behind using the Autoregressive Integrated Moving Average (ARIMA)
model for stock price prediction is rooted in the desire to understand and forecast stock price
movements.
 Risk Management: Accurate stock price predictions are crucial for managing risk in
investment portfolios.
 Trading Strategies: Stock traders use ARIMA predictions to inform their trading strategies.
ARIMA models can provide insights into potential short-term price movements, helping
traders make timely decisions for profit.
 Quantitative Analysis: ARIMA models offer a quantitative approach to stock price
forecasting.
 Technical analysts often use ARIMA predictions to complement their chart-based analyses.
ARIMA can provide additional data points to either confirm or challenge their technical
indicators and patterns.

1.2 PROBLEM DEFINITION

 Stock price prediction is crucial for investors, analysts, and financial institutions seeking to
make informed decisions. The stock market is highly volatile, impacted by various factors
such as economic indicators, market sentiment, and geopolitical events. Traditional statistical
models often struggle to capture these dynamics accurately over time, leading to limited
predictive success.
 The goal of this project is to design and implement a robust stock price prediction system
using the ARIMA model as a foundation, enhanced with machine learning techniques. The
ARIMA model will initially analyze historical time-series data, accounting for trends,
seasonality, and autocorrelations in stock prices. Further, machine learning techniques will be

12
 employed to improve prediction accuracy by analysing other market indicators or news
sentiment that may influence stock price.

2. LITERATURE SURVEY

Study/Source Focus Findings

"Hybrid ARIMA ARIMA + Random Forest Combining ARIMA with
and Machine Dataset: S&P 500 Index Random Forest improved
Learning Approach short-term forecast accuracy
for Stock Price by 15%.
Prediction"
(John Doe et al., 2020)
"Enhancing ARIMA ARIMA + LSTM LSTM layers captured non-
for Stock Prediction Dataset: Apple Inc. stock data linear patterns ARIMA
Using Deep missed, reducing MSE by
Learning 20%.
Techniques"
(Jane Smith et al., 2019).

"Machine Learning ARIMA + Support Vector SVM effectively

and Time Series Machine (SVM) complemented ARIMA for
Analysis for Stock Dataset: NASDAQ Data volatile market days,
Market Predictions" increasing prediction
accuracy in high-volatility
stocks.
(Kumar et al., 2020)

"Financial ARIMA + Feedforward Neural Neural network layers

Forecasting Network helped correct ARIMA’s
Using ARIMA Dataset: Indian Stock Exchange seasonal lag, achieving 12%
and Neural Data improvement in accuracy.
Networks"
(Lee & Lee, 2021)

"A Comparative ARIMA + Gradient Boosting Gradient Boosting provided

Analysis of ARIMA a stronger baseline than
and Hybrid Models Dataset: Dow Jones Industrial standalone ARIMA, with
for Stock Average hybrid model outperforming
Forecasting" ARIMA by 10% in R^2
values.

13
 ( Patel & Kothari, 2018).

Table 2: Literature Survey

14
 CHAPTER 3

SOFTWARE REQUIREMENT AND SPECIFICATION

INTRODUCTION

3.1.1 Project Scope

 Research and Analysis

 Conduct a literature review of traditional time-series methods (like ARIMA) and hybrid
approaches integrating machine learning.

15
  Identify common limitations of ARIMA in handling complex patterns and potential
solutions via machine learning algorithms.

 Data Collection and Preprocessing

 Gather historical stock price data (e.g., daily closing prices) from reliable sources (e.g.,
Yahoo Finance, Alpha Vantage).
 Preprocess data, addressing missing values, outliers, and preparing it for time series
analysis.

 Model Selection and Training

 Implement an ARIMA model to capture linear trends and seasonality in stock price data.
 Integrate machine learning algorithms (e.g., LSTM, Random Forest, Gradient Boosting)
to capture non-linear trends and patterns.
 Conduct training and optimization on both the ARIMA model and machine learning
models for hybrid forecasting.

 Testing and Validation

 Evaluate model accuracy using metrics such as Mean Squared Error (MSE), Root Mean
Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE).
 Perform cross-validation to assess the model’s stability and generalization across various
time periods.

 User Interface

 Develop a simple user interface for end-users to input stock symbols and receive
forecasted price predictions.
 Provide data visualizations of historical trends, ARIMA predictions, and machine
learning corrections.

User Classes and Characteristics

User Class Description Characteristics

Investors Individuals or firms - Knowledge of

seeking to predict stock financial markets
trends for investment - Interested in
decisions. long-term and
short-term stock
predictions

16
 - Intermediate
tech skills
Traders Professionals - High-frequency
making frequent trading needs
trades based on - Demands real-
market trends, time predictions
seeking precise, - Familiar with
short-term stock analysis
predictions. tools.

Data Analysts Analysts who - Strong analytical

research financial skills
trends and require - Requires raw data
stock data for and predictions
pattern analysis and - Uses various
reporting. models and
comparisons
ML Researchers Machine learning - Expertise in ML
specialists techniques
exploring ARIMA - Focus on model
and hybrid models performance and
in financial time- accuracy
series forecasting. - Seeks continuous
model
improvement
Developers Engineers - Proficient in coding
implementing stock and API integration
prediction systems - Focus on system
or integrating APIs stability
- Interested in
for trading
automation
platforms or capabilities
applications.

Financial Professionals - Basic

Advisors advising clients on understanding of
investment based predictive models
on model - Requires easy-to-
predictions. interpret insights
- Client-oriented

Business Decision-makers - High-level

Managers using stock understanding of
forecasts for prediction
business planning outcomes
and investment - Interested in

17
 strategies. summarized
insights
- Minimal technical
focus

Table 3.1.2 User Classes and Characteristics

18
 1.1 FUNCTIONAL REQUIREMENT

 Data Acquisition and Processing

 The system should retrieve historical stock price data via the Yahoo Finance API.
 It should support data retrieval for multiple stock symbols as specified by the user.
 Data processing functions must handle missing values, data transformations, and
normalization as required by the ARIMA model.

 Model Training and Selection

 The system should allow the selection of ARIMA model parameters (p, d, q) and offer
automated model selection options, such as grid search or Akaike Information Criterion
(AIC) for optimal parameters.
 It should support training the ARIMA model on historical data for a selected stock
symbol.
 The system should save and load trained models for reuse.

 Forecasting Functionality

 Users should be able to select a forecast range (e.g., 1 day, 7 days, 30 days).
 The system should generate stock price forecasts based on the ARIMA model for the
specified range.
 The system must provide options for viewing predictions as a time-series chart, with
overlayed actual and predicted values.

 Performance Metrics and Evaluation

 The system should calculate and display performance metrics for the ARIMA model,
such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean
Squared Error (RMSE).
 It should allow comparison between predicted and actual values over a defined period.

 User Interface and Interaction

 The system should offer a user-friendly interface where users can input stock symbols,
select parameters, view predictions, and interpret results easily.
 The system should provide an option to download predictions and historical data in CSV
format.

 Real-Time Alerts and Notifications

 The system should allow users to set specific conditions (e.g., price threshold) for
receiving notifications or alerts.
 It should provide notifications for major deviations between predicted and actual values,
if relevant.

19
1.2 EXTERNAL INTERFACE REQUIREMENTS

3.3.1 User Interfaces

Web Interface:

o Responsive design for accessibility on desktops, tablets, and mobile devices.

o User-friendly navigation with clear menus for projects, tasks, and user profiles.
o Dashboard displaying project summaries, task lists, and notifications.

Forms:

o Input forms for project creation, task assignment, and user registration/login.
o Modal dialogs for comments and file uploads.

3.3.2 Hardware Interfaces

Server Requirements:

o A cloud server to host the application, capable of handling multiple user requests.

User Devices:

o The application should be compatible with standard hardware (desktops, laptops,

tablets, and smartphones).

3.3.3 Software Interfaces

Database Interface:

o appropriate APIs for database interactions.

Web Frameworks:

20
 o Integration with Express.js for server-side logic and API handling.
o flask for the frontend user interface and state management.

3.3.4 Communication Interfaces

API Communication:

o RESTful APIs to facilitate communication between the frontend and backend.

Real-time Collaboration:

o Web Socket or similar technology for real-time updates and notifications among
users.

1.3 NON FUNCTIONAL REQUIREMENTS

3.4.1 Performance Requirements

 Latency: The system should respond to user queries within 1-2 seconds. Data retrieval
from Yahoo API should complete within 3-5 seconds, given a stable network.
 Throughput: The system should handle multiple concurrent users without significant
performance degradation, supporting at least 100 simultaneous users.
 Data Processing Speed: Data processing and model training should be optimized to
complete within a reasonable timeframe (e.g., under 2 minutes for training on 1 year of
data).
 Scalability: The system should be scalable, with a design that allows it to handle
additional stock symbols, datasets, and new data sources without compromising
performance.
 Fault Tolerance: In case of external API failure (e.g., Yahoo API downtime), the system
should retry the request up to three times before notifying the user of a connection issue.
21
3.4.2 Safety Requirements

 Data Handling Safety: The system should ensure safe handling of data inputs, especially
from external sources (e.g., Yahoo API), to prevent incorrect data affecting the
predictions.
 Data Backup: User preferences, model configurations, and saved models should be
backed up periodically to prevent data loss during updates or crashes.
 Error Handling: The system should provide informative error messages to users and
handle unexpected inputs gracefully, preventing crashes and ensuring the program
continues to operate normally.

3.4.3 Security Requirements

 Data Privacy: Ensure no personal identifiable information (PII) is collected unless

absolutely necessary, and implement safeguards to maintain user data privacy.
 API Security: Secure API access with unique API keys and implement encryption for
data in transit to protect communication with external services.
 Access Control: Implement access control to prevent unauthorized modifications to
model parameters or predictions. Users should have access only to their own models and
predictions.
 Audit and Logging: Maintain logs for user activity, particularly for data retrieval, model
training, and prediction functions, with proper data encryption to maintain security.

3.4.4 Software Quality Attributes

 Reliability: Ensure a stable environment where all functionalities, from data retrieval to
prediction, perform accurately under various conditions, with an uptime target of 99%.
 Usability: The user interface should be intuitive, with clear instructions, tooltips, and
easy navigation to enable both novice and expert users to operate the system without
extensive training.
 Maintainability: The system codebase should be modular and documented thoroughly to
allow for updates, including changes in API endpoints or the addition of new algorithms,
with minimal effort.

22
 Portability: The system should be compatible with different operating systems and web
browsers, offering flexibility for users to access it from various platforms.
 Extensibility: The design should allow for adding other forecasting models (e.g., LSTM,
Prophet) in the future, without requiring substantial rework of the system architecture.

3.5 SYSTEM REQUIREMENTS

3.5.1 Software Requirements

 Operating System: Compatible with Windows, macOS, or Linux environments for

flexibility in deployment.
 Programming Language: Python (preferably version 3.7 or higher) for implementing
machine learning models and data handling.
 Libraries and Frameworks:
 Machine Learning: scikit-learn and statsmodels for implementing ARIMA and other
prediction models.
 Data Processing: pandas, numpy for data manipulation and pre-processing.
 API Integration: yfinance or Yahoo Finance API for real-time stock data retrieval.
 Data Visualization: matplotlib and seaborn for visualizing stock trends and prediction
results.
 IDE/Code Editor: PyCharm, VSCode, or Jupyter Notebook for development and testing.
 Database Management: PostgreSQL, MongoDB, or SQLite depending on data volume
and project scale.
 Version Control: Git for source code management.
 Deployment: Docker or cloud environment like AWS, Azure, or Google Cloud for
scalability and remote access.

3.5.3 Hardware Requirements

23
Hardware Requirements

 Processor: Minimum quad-core processor (Intel i5 or equivalent) for handling data

processing and model training.
 Memory: At least 8 GB of RAM for smooth operation; 16 GB or higher is recommended
for handling larger datasets and training iterations.
 Storage: Minimum 100 GB SSD to accommodate both the software environment and the
database, with room for future data growth.
 GPU (Optional): Not mandatory for ARIMA, but a basic GPU like NVIDIA GTX 1050
or equivalent would be beneficial if deep learning models are planned for future
implementation.
 Network: High-speed internet connection for consistent API communication with Yahoo
Finance and fast data retrieval.

Table. 3.5.3 Hardware requirements

3.6 ANALYSIS MODELS: SDLC MODEL TO BE APPLIED

For a Stock Price Prediction System using an ARIMA model and Yahoo API, selecting an
appropriate System Development Life Cycle (SDLC) model is essential to accommodate the
complexity and dynamic nature of financial forecasting. Given the evolving requirements and
need for iterative refinement, here are some viable SDLC models:

24
1. Agile Model

 Overview: Agile focuses on iterative, incremental development with flexibility to adapt

to changing requirements. It is particularly well-suited for stock prediction projects where
feedback loops are essential for refining accuracy and adapting to market changes.
 Advantages:
o Continuous Improvement: Regular iterations allow for model tuning and
integration of new data.
o User Feedback: Each sprint can be reviewed with stakeholders, helping align the
system with user expectations for data handling, prediction intervals, and
accuracy.
o Risk Reduction: Risks are identified and addressed early with each iteration,
which is beneficial in a financial prediction context where accuracy is critical.

2. Incremental Model

 Overview: This model divides development into manageable increments, delivering

portions of the system at each stage. This is advantageous when a base predictive model
can be built initially, and additional features or optimizations are integrated
incrementally.
 Advantages:
o Early Delivery of Core Functionality: A basic ARIMA model for stock price
predictions can be deployed early for testing while other features are added.
o Scalable Additions: New features such as data visualization, alert notifications,
or prediction model refinements can be implemented incrementally.
o Flexibility with Development: Enables ongoing improvements and adjustments
without impacting the system's core functionality.

3. Waterfall Model

 Overview: The Waterfall model is a linear, sequential SDLC approach that may be
suitable if the project’s requirements are well-defined and less likely to change. While

25
 less flexible, it may work if stakeholders have clear and fixed expectations for the stock
prediction model and output.
 Advantages:
o Clear Phases: Each stage is planned out, making it straightforward to track
progress and manage resources.
o Less Rework: Since requirements are well-defined at the start, this model reduces
the likelihood of substantial changes.
o Ease of Documentation: Since each phase is completed before moving to the
next, thorough documentation can be maintained, which is beneficial for projects
requiring regulatory compliance or detailed audits.

Recommended SDLC Model: Agile Model

Given the complexity and iterative nature of stock price prediction, Agile is the most suitable
model for this project. This approach allows for continuous integration and testing, essential for
tuning the ARIMA model and incorporating real-time data through the Yahoo API. Agile’s
flexibility supports ongoing improvements, essential as model performance and user
requirements evolve with feedback and market data.

26
 3.7 SYSTEM IMPLEMENTATION PLAN

phase task Timeline Expected Outcome

1. Project Proposal - Define project Week 1 Approved project

goals and proposal
objectives
- Outline scope and
feasibility
2. Requirements Identify functional Week 2-3 Documented
Analysis & non-functional requirements and use
requirements cases

- Define user
requirements
3. System Design - Design system Week 4 System architecture and
data flow diagram
architecture

- Plan API Data acquisition

integration specifications
4. Data - Integrate Yahoo Week 5 Stock data collected
Collection & API to fetch and pre-processed for
Preprocessing historical data analysis
- Clean and Ready-to-use dataset
prepare data for
analysis
5. Model - Develop and Week 6-7 Trained ARIMA
Development configure model with optimal
ARIMA model parameters
- Test and Accuracy benchmarks
validate model and validation results
accuracy
6. Backend - Implement Week 8 Backend server
Development backend for supporting data and
prediction model integration
handling
- Set up data Real-time data
retrieval API retrieval API
from Yahoo
7. Frontend - Design user Week 9 Functional and user-
Development interface for

27
 displaying friendly UI
predictions
8. Integration & - Integrate Week 10 Fully integrated
Testing frontend, system
backend, and
model
- Conduct unit System passes
and integration required functional
testing tests
9. Project - Write technical Week 11 Comprehensive
Documentation and user project documentation
documentation
10. Final Report - Prepare final Week 12 Project report and
& Presentation report & presentation for
presentation evaluation

Table 3.7: System Implementation plan

28
 CHAPTER 4
SYSTEM DESIGN

29
4.1 System Architecture

Fig. 4.1 Architecture diagram

30
4.2 Data flow Diagram

Level-0 DFD

Fig. 4.2.1. Level-0 DFD

31
Level-1 DFD

Fig. 4.2.2. Level-1 DFD

32
4.3 Entity Relationship Diagrams

33
4.4 UML Diagrams

4.4.1 Use-case Diagram

Fig. 4.4.1 Use-case Diagram

34
 4.4.2 Class Diagram

Fig. 4.4.2 Use-case Diagram

35
 4.4.3 Activity Diagram

Fig. 4.4.3 Activity Diagram

36
4.4.4 sequence Diagram

Fig. 4.4.4 Sequence Diagram

37
 CHAPTER 05

OTHER SPECIFICATION

38
5.1 Advantages

 Accurate Short-Term Forecasting

 Real-Time Data Access
 Cost-Effective
 User-Friendly and Customizable
 Automation of Analysis
 enhanced Market Insights
 Scalable and Flexible system

5.2 Limitations

 Limited to Linear Patterns

 Short-Term Prediction Constraints

 Sensitivity to Market Volatility

 Dependency on Data Quality

 Limited Adaptability to External Influences

 High Maintenance and Model Updating

 Potential API Limitations

5.3 Applications

 Financial Markets and Trading

 Portfolio Management and Asset Allocation
 Investment Advisory Services
 Risk Management and Hedging
 Corporate Strategy and Financial Planning
 Personal Finance and Retail Investment
 Data Science and AI Development

39
 CHAPTER 6
CONCLUSIONS & FUTURE WORK

40
 Regarding the issue of stock price forecasting, many scholars are still studying in this area,
and using time series forecasting theory is feasible and effective for stock price
forecasting. As this article uses ARIMA model and BP neural network model to predict
the closing price of stocks. The empirical results show that these two models can predict
the future stock prices more accurately, and show that short-term forecasting of stock
prices is feasible and effective. However, this paper also has shortcomings. First, only
static forecasting is used, and it is still single-step forecasting. In fact, both time series
methods and neural networks can achieve multi-step forecasting. In addition, stock prices
are dynamic and continuous. Second, this paper only evaluates the fitting effect of the
prediction model from the relative error. It will be better if combined with the trend of
future stock price changes. Of course, stock price prediction itself does not form a
complete investment decision. At least it requires effective risk assessment and
corresponding risk control methods.

In short, this paper uses the historical closing price of stocks as time series data to
construct an ARIMA model and a BP neural network model, and make short-term
forecasts of the future stock opening prices. The forecasting effect is relatively ideal. The
two models are feasible and effective for short-term forecasting of stock price data.

 Enhanced Feature Engineering: Including additional external factors, such as economic

indicators, market sentiment, and global news, could improve the robustness of
predictions. Integrating sentiment analysis from news sources and social media using
natural language processing (NLP) could yield valuable insights.
 Real-time Prediction and Anomaly Detection: Enhancing the system for real-time
analysis and anomaly detection would allow it to issue alerts for unusual price behaviors,
aiding traders in responding quickly to market changes.
 Incorporating Risk Metrics: Developing risk analysis metrics alongside price prediction
would provide investors with a comprehensive understanding of not only probable price
movements but also associated risks.
 User-friendly Interface Development: Building a graphical interface with customizable
parameters and visualization tools for ease of use by financial analysts and retail
investors would make the system more accessible to non-technical users.

41
References:

1. Ashish Sharma, Dinesh Bhuriya, Upendra Singh. "Survey of Stock Market Prediction

Using Machine Learning Approach", ICECA 2022.

2. 2. Loke.K.S. “Impact Of Financial Ratios And Technical Analysis On Stock Price

Prediction Using Random Forests”, IEEE, 2021.

3. 3. Xi Zhang1, Siyu Qu1, Jieyun Huang1, Binxing Fang1, Philip Yu2, “Stock Market

Prediction via Multi-Source Multiple Instance Learning.” IEEE 2021.

4. 4. VivekKanade, BhausahebDevikar, SayaliPhadatare, PranaliMunde, ShubhangiSonone.

“Stock Market Prediction: Using Historical Data Analysis”, IJARCSSE 2022.

5. 5. SachinSampatPatil, Prof. Kailash Patidar, Asst. Prof. Megha Jain, “A Survey on Stock

Market Prediction Using SVM”, IJCTET 2021.

6. 6. https://www.cs.princeton.edu/sites/default/files/uploads/Saahil_magde. pdf

7. 7. Hakob GRIGORYAN, “A Stock Market Prediction Method Based on Support Vector

Machines (SVM) and Independent Component Analysis (ICA)”, DSJ 2021.

8. 8. RautSushrut Deepak, ShindeIshaUday, Dr. D. Malathi, “Machine Learning Approach

In Stock Market 9. Prediction”, IJPAM 2022.

42
43