Chhatrapati Shahu Ji Maharaj University
Project Synopsis
on
Stock Pre - Prediction Using Machine Learning
Submitted in Partial Fulfillment of the Requirements for the
Degree of
Bachelors of Computer Application
By
Arpit Awasthi, 22015001814
Anuradha, 22015001799
Under the Supervision of
Dr. Amit Yadav
Associate professor
PSIT College of Higher Education
Kanpur-Agra-Delhi National Highway – 2, Bhauti, Kanpur
(2024-2025)
0
ABSTRACT
The stock market is highly volatile and influenced by various factors such as economic conditions,
political events, and investor sentiment. Predicting stock prices is a complex task due to the market's
dynamic nature. This project focuses on Stock Price Prediction using Machine Learning, where
historical stock data is analysed to forecast future prices. Various machine learning techniques,
including Linear Regression, Random Forest, Support Vector Machines (SVM), and Long Short-
Term Memory (LSTM) networks, are implemented to enhance prediction accuracy. The project aims
to provide investors with a data-driven approach to making informed financial decisions.
1
CONTENTS
Page no.
1. Introduction 3
1.1 Objective 3
1.2 Scope 3
2. System Requirement
2.1 Software Requirement 4
2.2 Hardware Requirement 4
3. Modules 5
3.1 Modules Name 5
3.2 Modules Description 5
4. Application Architecture 6
5. Technology Used 7
6. References 8
2
1. INTRODUCTION
1.1 Objective
The primary objective of this project is to develop a machine learning model that can predict stock
prices based on historical data. The model aims to:
Utilize past stock market data to predict future trends.
Compare different ML algorithms for accuracy and reliability.
Provide insights into market behaviour using technical indicators and statistical analysis.
Assist investors in making informed financial decisions by offering data-backed predictions.
1.2 Scope
This project focuses on short-term and long-term stock price prediction using machine learning
models. The scope includes:
Data Collection: Fetching stock data from financial APIs (Yahoo Finance, Alpha Vantage,
etc.).
Data Preprocessing: Cleaning and handling missing values.
Feature Engineering: Analyzing stock market indicators such as moving averages, RSI,
MACD, and Bollinger Bands.
Model Development: Implementing machine learning models like Linear Regression,
Random Forest, and LSTM.
Evaluation: Measuring model accuracy using metrics like Mean Squared Error (MSE), Root
Mean Squared Error (RMSE), and R-Squared Score.
3
2. SYSTEM REQUIREMENTS
2.1 Software Requirements
Programming Languages: Python
Machine Learning Libraries: TensorFlow, Keras, Scikit-learn
Data Processing Libraries: Pandas, NumPy, Matplotlib, Seaborn
Web Frameworks: Flask / Django (for deployment)
Database: SQLite / MySQL (for storing historical stock data)
Development Tools: Jupyter Notebook, Google Colab, Anaconda
2.2 Hardware Resource Requirements
Processor: Intel i5/i7 or AMD Ryzen equivalent
RAM: Minimum 8GB (16GB recommended for deep learning models)
Storage: At least 50GB free space (For large datasets)
GPU: Recommended for deep learning models like LSTM (e.g. NVIDIA GTX 1060 or
higher)
4
3. MODULES
3.1 Modules Name
1. Data Collection Module
2. Data Preprocessing Module
3. Feature Engineering Module
4. Model Training & Prediction Module
5. Deployment Module
3.2 Modules Description
Data Collection Module: Fetches stock market data from Yahoo Finance, Alpha Vantage, or
Quandl API. Collects historical stock prices, trading volume, and other financial indicators.
Stores data in a structured format for further processing.
Data Preprocessing Module: Cleans the dataset by handling missing values and outliers.
Normalizes numerical values to ensure consistent scale across different stock attributes. Splits
data into training and testing sets.
Feature Engineering Module: Extracts useful features such as moving averages, Bollinger
Bands, RSI (Relative Strength Index), MACD (Moving Average Convergence
Divergence).Converts time-series data into supervised learning format.
Model Training & Prediction Module: Implements various machine learning models.
Linear Regression for basic price trend [Link] Forest for reducing variance in
[Link] Vector Machines (SVM) for pattern recognition. Long Short-Term
Memory (LSTM) for deep learning-based time series forecasting. Trains and fine-tunes
models using historical data. Evaluates model performance using Mean Absolute Error
(MAE) and RMSE.
5
Deployment Module: Integrates the model into a web application using Flask. Displays real-
time stock predictions in a user-friendly dashboard. Allows users to input stock symbols and
get future predictions.
4. APPLICATION ARCHITECTURE
1. Data Collection & Ingestion: -The system gathers historical stock prices, technical indicators,
and external factors (news sentiment, macroeconomic data) from APIs like Alpha Vantage,
Yahoo Finance, or Quandl. The data is stored in a database (SQL, NoSQL) or a cloud-based data
lake.
2. Data Preprocessing & Feature Engineering: -Raw data is cleaned, normalized, and
transformed. Missing values are handled, and features like moving averages, RSI, and Bollinger
Bands are computed. Sentiment analysis can be performed on news headlines using NLP
techniques.
3. Model Selection & Training: -Machine learning models such as LSTMs, Random Forests, or
XGBoost are trained on historical data. Deep learning approaches like RNNs or Transformers can
be used for time-series forecasting. The model is trained using frameworks like TensorFlow or
PyTorch and evaluated using metrics like RMSE and R².
4. Prediction & Deployment: -Once trained, the model is deployed using Flask, FastAPI, or
Django as a REST API. The API receives real-time stock data, processes it, and generates
predictions. A front-end (React, Angular) provides a dashboard for users to view forecasts and
analytics.
5. Continuous Monitoring & Updating: -A monitoring system tracks model accuracy and
performance. Retraining pipelines using MLOps tools like MLflow or Kubeflow ensure model
updates based on new data. Cloud platforms (AWS, GCP, Azure) handle scalability.
6
5. TECHNOLOGY USED
5.1 Frontend:
HTML, CSS, JavaScript: Used for creating a responsive and user friendly interface.
[Link] / [Link]: Used for building an interactive front end experience.
5.2 Backend:
Python (flask/Django): Manages the server -side logic and API request for stock prediction.
REST API Development: Facilitates data communication between the frontend and backend.
5.3 Database:
SQLite/MYSQL: Stores historical stock data and user interactions.
MongoDB: Used for handling large unstructured datasets.
5.4 Framework:
Machine Learning: Scikit-learn, TensorFlow, Keras
Data Processing: Pandas, Numpy
Visualization: Matplotlib, Seaborn
Deployment Platforms: Heroku, AWS, Google cloud
7
6. REFERENCES
1. Stock pre-prediction using Python & Machine Learning: -
[Link]
2. Datasets for Stock Prediction: -
[Link] Finance API (Real-time & historical stock data): -
[Link]
2.2. Alpha Vantage API (Stock prices, indicators, news sentiment): -
[Link]
2.3. Quandl (Financial, economic & alternative data): - [Link]
2.4. Kaggle Datasets for Stock Market Prediction: -
[Link]
3. Stock Price Prediction using LSTM in Python: -
[Link]
python-734dd05fef8f
4. Tech Stack Recommendations: -
4.1. Backend (Data Processing & Model Training): -
4.1.1. Python (Pandas, NumPy, Scikit-learn, TensorFlow, PyTorch)
4.1.2. Flask/Django (If building a web-based app)
4.1.3. FastAPI (For API services)
4.1.4. PostgreSQL/MySQL/MongoDB (For storing predictions)
4.2. Frontend (Visualization & User Interface): -
4.2.1. [Link] (For interactive UI)
4.2.2. [Link] & [Link] (For stock data visualization)
4.2.3. Streamlit / Dash (For quick prototyping)