VISVESVARAYA TECHNOLOGICAL UNIVERSITY

“Jnana Sangama”, Belagavi - 590018, Karnataka, India

MINI PROJECT REPORT

On

“Stock Price Prediction Using Machine Learning”

Submitted in partial fulfillment of the requirements for the award of the Degree
BACHELOR OF ENGINEERING
In
ELECTRONICS AND COMMUNICATION ENGINEERING
By

SHIVANAND J SHINDHE 1DA21EC133

SIDDESH G M 1DA21EC139
VINAY C KADECHUR 1DA21EC166

Under the Guidance of

NITHYASHREE S
Asst. Prof. Dept of ECE,
Dr. AIT, Bengaluru –
560056

DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING

Dr. AMBEDKAR INSTITUTE OF TECHNOLOGY
MALLATHAHALLI, OUTER RING ROAD, BENGALURU – 560056
2023-24
 Dr. AMBEDKAR INSTITUTE OF TECHNOLOGY
Mallathahalli, Bengaluru – 560056
Department of Electronics & Communication Engineering

Certificate
Certified that the project work entitled “Stock Price Prediction Using Machine
Learning ”, carried out by Shivanand J Shindhe, bearing USN: 1DA21EC133, Siddesh G
M, bearing USN: 1DA21EC139 and Vinay C Kadechur, bearing USN: 1DA21EC166,
bonafide students of Dr. Ambedkar Institute of Technology, Bangalore – 560056 in
partial fulfillment for the award of Bachelor of Engineering in Electronics and
Communication Engineering of the Visvesvaraya Technological University, Belagavi
during the year 2022–2023. It is certified that all the corrections/suggestions indicated for
Internal Assessment have been incorporated in the Report deposited in the departmental
library. The Mini project report has been approved as it satisfies the academic requirements.

Signature of the Guide Signature of the HOD Signature of the Principal

(Nithyashree S) (Dr. Ramesh S) (Dr. C. Nanjundaswamy)

External Viva
Name of the Examiners Signature with Date

1.

2.
 Dr. AMBEDKAR INSTITUTE OF TECHNOLOGY
Mallathahalli, Bengaluru – 560056

Department of Electronics and Communication Engineering

Declaration
We, SHIVANAND J SHINDHE, bearing USN : 1DA21EC133, SIDDESH G M bearing
USN :1DA21EC139, and VINAY C KADECHUR, bearing USN:1DA21EC166, hereby
declare that, the project work entitled “Stock Price Prediction Using Machine Learning”
is independently carried out by us at Department of Electronics and Communication Engineering,
Dr. Ambedkar Institute of Technology, Bengaluru-560056, under the guidance of
NITHYASHREE S, Assistant Professor, Department of Electronics and Communication
Engineering, Dr. Ambedkar Institute of Technology. The Project work is carried out in partial
fulfillment of the requirement for the award of degree of Bachelor of Engineering in Electronics and
Communication Engineering during the academic year 2023- 2024.

Place: Bengaluru Name & Signature of students

Date: SHIVANAND J SHINDHE

SIDDESH G M
VINAY C KADECHUR
 ACKNOWLEDGEMENT

The satisfaction that accompanies the successful completion of this mini project would be
complete only with the mention of the people who made it possible, whose support rewarded
our effort with success.

We are grateful to Dr. Ambedkar Institute of Technology for its ideals and its inspirations
for having provided us with the facilities that have made this mini project a success.

We are grateful to our Principal Dr. C Nanjundaswamy, Dr. Ambedkar Institute of

Technology, who gave a continuous support and provided us comfortable environment to
work in.

We would like to express our sincere thanks to Dr. Ramesh S, Professor and Head,
Department of Electronics and Communication Engineering, Dr. Ambedkar Institute of
Technology for his support. We pay out profound gratefulness and express our deepest
gratitude to our mini project guide Nithyashree S, Assistant Professor, Department of
Electronics and Communication Engineering for the suggestions and guidance.

We are thankful to our mini project coordinators Dr. Divya A, Assistant Professor, Dept. of
Electronics and communication and Sangeetha N, Assistant Professor, Dept. of Electronics
and communications for their advice, supervision and guidance throughout the course of the
project.

It is our pleasure to acknowledge the cooperation extended by teaching staff and non-
teaching staff members of Dept. of Electronics and Communication Engineering, Dr.
Ambedkar Institute of Technology for the encouragement during project work. Finally, it
gives immense pleasure to acknowledge the cooperation extended by family members,
friends for the encouragement during this Project Work.

SHIVANAND J SHINDHE
SIDDESH G M
VINAY C KADECHUR
 ABSTRACT

Stock price prediction is a critical task in financial markets, providing insights that can drive investment
strategies and decision-making. This report explores the application of machine learning techniques to
forecast stock prices, leveraging historical market data and a range of technical and macroeconomic
indicators. The objective is to build a predictive model that accurately estimates future stock prices,
thereby enabling investors to make informed trading decisions.

In this study, various machine learning algorithms are employed, including Linear Regression, Random
Forest, and Long Short-Term Memory (LSTM) networks, each chosen for their specific strengths in
handling time series data and capturing complex patterns. The process begins with extensive data
preprocessing to clean and normalize the data, followed by feature engineering to extract meaningful
predictors. These features include moving averages, trading volumes, and momentum indicators, which
are critical in understanding stock price movements.

The models are trained and evaluated using historical stock data, with performance assessed through
metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared values.
Back testing is conducted to validate the model's predictions against actual market performance, ensuring
robustness and reliability.

Results indicate that machine learning models can effectively predict stock prices, with the LSTM
demonstrating superior performance due to its ability to capture temporal dependencies and trends. The
findings underscore the potential of machine learning as a powerful tool for financial forecasting,
significant advantages over traditional statistical methods.
 INDEX

Chapter 1 Introduction Pg.No.

1.1 Introduction 1
1.2 Problem Statement 2

Chapter 2 Literature Survey

2.1 Literature 3-4
2.2 Objectives 4-5

Chapter 3 Proposed Work

3.1 Methodology 6-7
3.2 Flow chart 8-12
3.3 Requirements and Specifications 13-14
3.4 Implementation 14-17

Chapter 4 Result
4.1 Results 18-20
4.2 Application 21-22
4.3 Advantages 23-24

Chapter 5 Conclusion and Future scope

5.1 Conclusion 25
5.2 Future Scope 25-27
5.3 Reference 28-29
 LIST OF FIGURES

List No. Title Page No.

3.1 Flow chart 8

3.2 Code snippet 15-17

3.3 Stock data values 18

3.4 Wave form 19-20

CHAPTER 1
INTRODUCTION
1.1 INTRODUCTION

Stock price prediction has long been a focal point in the financial industry, driven by the goal of
maximizing investment returns and managing risk. Accurately forecasting stock prices enables
investors, traders, and financial analysts to make informed decisions that can lead to significant
financial gains. Traditional methods of stock price prediction often rely on fundamental analysis,
technical analysis, or a combination of both. However, the inherent complexity and dynamic
nature of financial markets pose substantial challenges to these conventional approaches.

In recent years, the advent of machine learning has introduced new possibilities for stock price
prediction. Machine learning, a subset of artificial intelligence, employs algorithms that can learn
patterns from vast amounts of data and make predictions based on those patterns. These
algorithms have the potential to uncover intricate relationships within financial data that may be
imperceptible to traditional statistical methods.

This report explores the application of machine learning techniques to predict stock prices,
utilizing a comprehensive dataset comprising historical stock prices, technical indicators, and
macroeconomic variables. The study aims to build and evaluate models that can accurately
forecast future stock prices, thereby providing valuable insights for investment strategies.

The report is structured as follows: we begin with a review of related work in stock price
prediction using machine learning, highlighting the advantages and limitations of various
approaches. Next, we describe the methodology, including data collection, preprocessing,
feature engineering, and model selection. The core of the report details the development and
evaluation of multiple machine learning models, such as Linear Regression, Random Forest, and
Long Short-Term Memory (LSTM) networks. Performance metrics and back testing results are
presented to assess the models' predictive capabilities.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 8

1.2 PROBLEM STATEMENT

The financial markets are characterized by their volatility and complexity, making accurate stock
price prediction a challenging yet highly sought-after goal. Traditional methods of forecasting
stock prices, which rely on fundamental and technical analysis, often fall short due to the
dynamic and multifaceted nature of market data. Investors and traders require more robust and
sophisticated tools to predict stock prices effectively, in order to make informed investment
decisions and develop profitable trading strategies.

In this context, machine learning offers a promising solution, with its ability to analyze large
datasets, identify hidden patterns, and make predictions based on historical trends and various
indicators. Despite its potential, the application of machine learning to stock price prediction
presents several challenges, including the need for high-quality data preprocessing, appropriate
feature selection, and the choice of suitable algorithms that can handle the intricacies of financial
time series data.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 9

CHAPTER 2
LITERATURE SURVEY
2.1 LITERATURE SURVEY

1. "Stock Price Prediction Using Machine Learning Techniques: A Review" by Rajiv Yadav et
al. (2016) - This comprehensive review paper provides an overview of various machine
learning techniques applied to stock price prediction. It covers traditional method such as
Linear Regression and Support Vector Machines, as well as more advanced techniques
including Artificial Neural Networks and Genetic Algorithms. The paper discusses the
strengths and limitations of each approach and offers insights into future research directions.

2. "Predicting Stock Prices Using Technical Analysis and Machine Learning" by Nitish
Veerendra et
al. (2017): - Veerendra et al. explore the effectiveness of combining technical analysis
indicators with machine learning algorithms for stock price prediction. The study evaluates
popular technical indicators such as Moving Averages, Relative Strength Index (RSI), and
Bollinger Bands, along with machine learning algorithms including Random Forest and
Gradient Boosting. The findings suggest that incorporating technical indicators can enhance
the predictive accuracy of machine learning models

3. "Stock Price Prediction Using LSTM, RNN and CNN-SVM" by Debaditya Roy et al. (2019):-
. compare the performance of different deep learning architectures, including Long
Short-Term
Memory (LSTM), Recurrent Neural Networks (RNN), and Convolutional Neural Network
(CNN),
combined with Support Vector Machines (SVM) for stock price prediction.

4. "A Survey on Stock Price Prediction Using Machine Learning Techniques" by C. R.

Vijayalakshmi et al. (2020): - This survey paper provides a comprehensive of recent
advancements in stock price prediction using machine learning techniques. It discusses
various data sources, feature selection methods, and machine learning algorithms
commonly employed in the field. Additionally, the paper highlights emerging trends and

Dept. of ECE, Dr.AIT, Bangalore-56 Page 10

 challenges in stock market prediction and offers recommendations for future
research directions.
5. “Stock Market Prediction using Machine Learning” by Subham Kumar Gupta1, Dr. Bhuvana J2, Dr.
M N Nachappa3 The literature survey on stock price prediction using LSTM, ARIMA, and
other methods is extensive and highlights several critical studies and methodologies. The
application of moving averages, ARIMA models, and LSTM neural networks has been
thoroughly explored by researchers to predict stock prices. Moving averages provide a
simple yet effective way to smooth out price data to identify trends. ARIMA models are
useful for short-term predictions and are valued for their ability to handle time-series data
with trends and seasonality. LSTM networks, part of the recurrent neural network family,
are adept at capturing the long-term dependencies in sequential data, making them suitable
for stock market predictions where historical prices significantly influence future prices.
Combining these models, as noted in recent studies, can capture both linear and non-linear
patterns in stock price data, thus improving the forecast accuracy. The literature also
discusses various factors influencing stock prices, such as market sentiment, news events,
and economic indicators. Despite advancements, predicting stock prices remains challenging
due to market volatility and the influence of unpredictable external factors. Continuous
improvements and hybrid model developments are ongoing to enhance prediction accuracy
and reliability.

2.2 OBJECTIVES

1. Data Collection and Preprocessing:

o To gather comprehensive historical stock price data from reliable sources.
o To preprocess the collected data by handling missing values, outliers, and normalization
to ensure quality and consistency.
2. Feature Engineering and Selection:
o To identify and engineer relevant features from the raw data, including technical
indicators (e.g., moving averages, RSI, MACD) and macroeconomic variables (e.g.,
interest rates, GDP growth).
o To select the most significant features that contribute to the predictive power of the
model.
3. Model Development:

Dept. of ECE, Dr.AIT, Bangalore-56 Page 11

 o To explore various machine learning algorithms such as Linear Regression, Support
Vector Machines (SVM), Random Forests, and deep learning models like Long Short-
Term Memory (LSTM) networks.
o To develop and train these models on the preprocessed data, optimizing their parameters
to achieve the best performance.
4. Model Evaluation and Validation:
o To evaluate the performance of the developed models using appropriate metrics such as
Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared.
o To validate the models through back testing on historical data, assessing their predictive
accuracy and robustness in real-world scenarios.
5. Comparison and Analysis:
o To compare the performance of different machine learning models in predicting stock
prices.
o To analyze the strengths and weaknesses of each model, providing insights into their
suitability for various market conditions.
6. Implementation and Visualization:
o To implement a practical and scalable pipeline for real-time stock price prediction.
o To develop interactive visualization tools for presenting the predictions and comparing
them with actual stock prices, aiding in decision-making processes.

7. Conclusion and Future Work :

o Summarize the findings and highlight the effectiveness of LSTM in capturing temporal
Suggest improvements for future research, such as incorporating sentiment analysis,
exploring advanced deep learning architectures, and enhancing model interpretability.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 12

CHAPTER 3
PROPSED WORK
3.1 METHODOLOGY

Data Collection:

• Historical Stock Price Data: Obtain historical stock prices for the target stocks from reliable
financial databases such as Yahoo Finance, Google Finance, or Bloomberg. Data should
include daily closing prices, opening prices, high and low prices, and trading volumes.
• Technical Indicators: Compute technical indicators such as Moving Averages (MA),
Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD),
Bollinger Bands, and others based on the historical price data.
• Macroeconomic Variables: Collect relevant macroeconomic data such as interest rates,
GDP growth rates, inflation rates, and unemployment rates from economic databases like the
Federal Reserve Economic Data (FRED) or the World Bank.

2. Data Preprocessing:

• Data Cleaning: Handle missing values through interpolation or imputation techniques.

Remove outliers using statistical methods or domain knowledge.
• Normalization: Normalize the data to a standard scale (e.g., using Min-Max scaling or Z-
score normalization) to ensure that all features contribute equally to the model training.
• Feature Engineering: Generate additional features that might help in prediction, such as
lagged returns, trading volume changes, and ratios of different technical indicators.
• Feature Selection: Use techniques like correlation analysis, Principal Component Analysis
(PCA), or feature importance from tree-based models to select the most relevant features for
the prediction task.

3. Model Development:

• Algorithm Selection: Choose a diverse set of machine learning algorithms for comparison,
including:
o Linear Regression: To establish a baseline model.
o Support Vector Machines (SVM): For handling high-dimensional feature spaces.
o Random Forests: For robust predictions and feature importance analysis.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 13

 o Long Short-Term Memory (LSTM) Networks: For capturing temporal
dependencies in time series data.
• Model Training: Split the dataset into training and testing sets (e.g., 80-20 split). Train each
model on the training set using cross-validation to ensure robustness and avoid overfitting.
• Hyperparameter Tuning: Perform hyperparameter optimization using techniques such as
Grid Search or Random Search to find the best-performing model parameters.

4. Model Evaluation:

• Performance Metrics: Evaluate the models using metrics such as Mean Absolute Error
(MAE), Root Mean Squared Error (RMSE), and R-squared to measure prediction accuracy.
• Back testing: Conduct back testing by applying the trained models to a portion of historical
data that was not used in training to simulate real-world performance and validate the model's
predictive power.

5. Implementation and Visualization:

• Prediction Pipeline: Develop a scalable pipeline for real-time stock price prediction,
including data ingestion, preprocessing, model inference, and output generation.
• Visualization Tools: Create interactive dashboards and visualizations using tools like
Matplotlib, Seaborn, or Plotly to display predicted stock prices alongside actual prices,
enabling easy comparison and analysis.

6. Comparison and Analysis:

• Model Comparison: Compare the performance of different models based on evaluation

metrics and visual inspection of predicted versus actual stock price trends.
• Strengths and Weaknesses: Analyze the strengths and weaknesses of each model,
discussing their suitability for various market conditions and their ability to generalize across
different stocks and time periods.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 14

3.2 FLOW CHART

Dept. of ECE, Dr.AIT, Bangalore-56 Page 15

3.2.1 Long Short-Term Memory
Long Short-Term Memory is an improved version of recurrent neural network designed by Hochreiter
& Schmidhuber.
A traditional RNN has a single hidden state that is passed through time, which can make it difficult for
the network to learn long-term dependencies. LSTMs model address this problem by introducing a
memory cell, which is a container that can hold information for an extended period.
LSTM architectures are capable of learning long-term dependencies in sequential data, which makes
them well-suited for tasks such as language translation, speech recognition, and time series forecasting.
LSTM Architecture:
The LSTM architectures involves the memory cell which is controlled by three gates: the input gate,
the forget gate, and the output gate. These gates decide what information to add to, remove from, and
output from the memory cell.
• The input gate controls what information is added to the memory cell.
• The forget gate controls what information is removed from the memory cell.
• The output gate controls what information is output from the memory cell.
This allows LSTM networks to selectively retain or discard information as it flows through the
network, which allows them to learn long-term dependencies.
The LSTM maintains a hidden state, which acts as the short-term memory of the network. The hidden
state is updated based on the input, the previous hidden state, and the memory cell’s current state.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 16

LSTM Working:

LSTM architecture has a chain structure that contains four neural networks and different memory
blocks called cells.
Information is retained by the cells and the memory manipulations are done by the gates. There are
three gates –
Forget Gate
The information that is no longer useful in the cell state is removed with the forget gate. Two
inputs xt (input at the particular time) and ht-1 (previous cell output) are fed to the gate and multiplied
with weight matrices followed by the addition of bias. The resultant is passed through an activation
function which gives a binary output. If for a particular cell state the output is 0, the piece of
information is forgotten and for output 1, the information is retained for future use. The equation for
the forget gate is:
𝑓𝑡=𝜎(𝑊𝑓⋅[ℎ𝑡−1,𝑥𝑡]+𝑏𝑓) ft=σ(Wf⋅[ht−1,xt]+bf)
where:

Dept. of ECE, Dr.AIT, Bangalore-56 Page 17

 • W_f represents the weight matrix associated with the forget gate.
• [h_t-1, x_t] denotes the concatenation of the current input and the previous
hidden state.
• b_f is the bias with the forget gate.
• σ is the sigmoid activation function.

Input gate
The addition of useful information to the cell state is done by the input gate. First, the information is
regulated using the sigmoid function and filter the values to be remembered similar to the forget gate
using inputs ht-1 and xt. . Then, a vector is created using tanh function that gives an output from -1 to
+1, which contains all the possible values from ht-1 and xt. At last, the values of the vector and the
regulated values are multiplied to obtain the useful information. The equation for the input gate is:
𝑖𝑡=𝜎(𝑊𝑖⋅[ℎ𝑡−1,𝑥𝑡]+𝑏𝑖) it=σ(Wi⋅[ht−1,xt]+bi)
𝐶^𝑡=𝑡𝑎𝑛ℎ(𝑊𝑐⋅[ℎ𝑡−1,𝑥𝑡]+𝑏𝑐)C^t=tanh(Wc⋅[ht−1,xt]+bc)
We multiply the previous state by ft, disregarding the information we had previously chosen to ignore.
Next, we include it∗Ct. This represents the updated candidate values, adjusted for the amount that we
chose to update each state value.
𝐶𝑡=𝑓𝑡⊙𝐶𝑡−1+𝑖𝑡⊙𝐶^𝑡Ct=ft⊙Ct−1+it⊙C^t
where
• ⊙ denotes element-wise multiplication
• tanh is tanh activation function

Output gate
The task of extracting useful information from the current cell state to be presented as output is done
by the output gate. First, a vector is generated by applying tanh function on the cell. Then, the
information is regulated using the sigmoid function and filter by the values to be remembered using
inputs ℎ𝑡−1ht−1and 𝑥𝑡xt. At last, the values of the vector and the regulated values are multiplied to
be sent as an output and input to the next cell. The equation for the output gate is:
𝑜𝑡=𝜎(𝑊𝑜⋅[ℎ𝑡−1,𝑥𝑡]+𝑏𝑜)ot=σ(Wo⋅[ht−1,xt]+bo)

Dept. of ECE, Dr.AIT, Bangalore-56 Page 18

Flow Chart:

Algorithm:
1. Initialize weights for each neuron: This step involves setting up the initial weights
for the LSTM network, which can be done using various initialization methods.
2. Read the training examples: Load the dataset that will be used to train the LSTM
network.
3. Training Loop: This loop iterates over each training example, passing it through
the network, and comparing the output to the expected result.
▪ If the output is correct, the algorithm checks if it has reached the end of the training
examples.
▪ If the output is incorrect, the weights are adjusted (typically using
backpropagation) to improve the network’s performance.
4. End: The training process concludes when all training examples have been processed
and the end of the list is reached.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 19

3.3 SPECIFICATIONS AND REQUIREMENTS

SOFTWARE REQUIREMENTS:

Programming Languages

• Python: Python is the primary programming language used for this project due to its
simplicity, extensive libraries, and strong community support. It is well-suited for data
analysis, machine learning, and visualization tasks.

Development Environment

• Jupyter Notebook: An interactive development environment that allows for easy data
exploration, visualization, and iterative model development. Jupyter Notebooks are
particularly useful for documenting the analysis process and sharing results.

Libraries and Frameworks

• Pandas: A powerful library for data manipulation and analysis. It provides data structures
like Data Frames to handle and preprocess large datasets.
• NumPy: A fundamental package for numerical computations in Python. It is used for array
operations and efficient numerical processing.
• Scikit-learn: A comprehensive library for machine learning in Python. It includes tools for
data preprocessing, feature selection, and implementing various machine learning
algorithms.

• TensorFlow/Keras: Deep learning libraries used for building and training neural networks,
including Long Short-Term Memory (LSTM) networks. Keras, as a high-level API,
simplifies the development of deep learning models.

• Matplotlib: A plotting library for creating static, animated, and interactive visualizations in
Python.

Integrated Development Environment (IDE)

Dept. of ECE, Dr.AIT, Bangalore-56 Page 20

 • Visual Studio Code: A lightweight, versatile code editor that supports Python development
with extensions for code linting, debugging, and version control.

Version Control

• Git: A version control system for tracking changes in the source code during software
development. It allows for collaboration and efficient management of code repositories.
• GitHub: A web-based platform that hosts Git repositories, facilitating version control,
collaboration, and project management.

Additional Tools

• APIs for Data Collection: APIs such as Yahoo Finance API, Alpha Vantage, or Quandl for
fetching historical stock price data and relevant financial information.

Computational Resources

• Local Machine: For initial development and testing, a personal computer with sufficient
processing power and memory can be used.

3.4 IMPLEMENTATION

1. Data Collection and Preprocessing:

o To gather comprehensive historical stock price data from reliable sources.
o To preprocess the collected data by handling missing values, outliers, and normalization
to ensure quality and consistency.
2. Feature Engineering and Selection:
o To identify and engineer relevant features from the raw data, including technical
indicators (e.g., moving averages, RSI, MACD) and macroeconomic variables (e.g.,
interest rates, GDP growth).
o To select the most significant features that contribute to the predictive power of the
model.
3. Model Development:
o To explore various machine learning algorithms such as Linear Regression, Support
Vector Machines (SVM), Random Forests, and deep learning models like Long Short-
Term Memory (LSTM) networks.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 21

 o To develop and train these models on the preprocessed data, optimizing their parameters
to achieve the best performance.

4. Model Evaluation and Validation:

o To evaluate the performance of the developed models using appropriate metrics such as
Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared.
o To validate the models through back testing on historical data, assessing their predictive
accuracy and robustness in real-world scenarios.
5. Comparison and Analysis:
o To compare the performance of different machine learning models in predicting stock
prices.
o To analyze the strengths and weaknesses of each model, providing insights into their
suitability for various market conditions.
6. Implementation and Visualization:
o To implement a practical and scalable pipeline for real-time stock price prediction.
7. Conclusion and Future work:

o Summarize the findings and highlight the effectiveness of LSTM in capturing temporal
o Suggest improvements for future research, such as incorporating sentiment analysis,
exploring advanced deep learning architectures, and enhancing model interpretability.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import yfinance as yf
start = '2012-01-01'
end = '2024-07-11'
stock = 'GOOG'

data = yf.download(stock, start, end)

data.reset_index(inplace=True)
data
ma_100_days = data.Close.rolling(100).mean()
plt.figure(figsize=(8,6))
plt.plot(ma_100_days, 'r')
plt.plot(data.Close, 'g')
plt.show()
ma_200_days = data.Close.rolling(200).mean()
plt.figure(figsize=(8,6))
plt.plot(ma_100_days, 'r')

Dept. of ECE, Dr.AIT, Bangalore-56 Page 22

plt.plot(ma_200_days,'b')
plt.plot(data.Close,'g')
plt.show()
data.dropna(inplace=True)
data_train = pd.DataFrame(data.Close[0: int(len(data)*0.80)])
data_test = pd.DataFrame(data.Close[int(len(data)*0.80): len(data)])
data_train.shape[0]
data_test.shape[0]
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler(feature_range=(0,1))
data_train_scale = scaler.fit_transform(data_train)
x = []
y = []

for i in range(100, data_train_scale.shape[0]):

x.append(data_train_scale[i-100:i])
y.append(data_train_scale[i,0])

x, y = np.array(x), np.array(y)
from keras.layers import Dense, Dropout, LSTM
from keras.models import Sequential
model = Sequential()
model.add(LSTM(units = 50, activation = 'relu', return_sequences = True,
input_shape = ((x.shape[1],1))))
model.add(Dropout(0.2))

model.add(LSTM(units = 60, activation='relu', return_sequences = True))

model.add(Dropout(0.3))

model.add(LSTM(units = 80, activation = 'relu', return_sequences = True))

model.add(Dropout(0.4))

model.add(LSTM(units = 120, activation = 'relu'))

model.add(Dropout(0.5))

model.add(Dense(units =1))
model.compile(optimizer = 'adam', loss = 'mean_squared_error')
model.fit(x,y, epochs = 50, batch_size =32, verbose =1)
model.summary()
pas_100_days = data_train.tail(100)
data_test = pd.concat([pas_100_days, data_test], ignore_index=True)
data_test_scale = scaler.fit_transform(data_test)
x = []
y = []

for i in range(100, data_test_scale.shape[0]):

x.append(data_test_scale[i-100:i])
y.append(data_test_scale[i,0])
x, y = np.array(x), np.array(y)
y_predict = model.predict(x)

Dept. of ECE, Dr.AIT, Bangalore-56 Page 23

scale =1/scaler.scale_
y_predict = y_predict*scale
y = y*scale
plt.figure(figsize=(10,8))
plt.plot(y_predict, 'r', label = 'Predicted Price')
plt.plot(y, 'g', label = 'Original Price')
plt.xlabel('Time')
plt.ylabel('Price')
plt.legend()
plt.show()
model.save('Stock Predictions Model.keras')

Dept. of ECE, Dr.AIT, Bangalore-56 Page 24

CHAPTER 4

RESULT

4.1 RESULT:
This section presents the results of the stock price prediction using the Long Short-Term Memory
(LSTM) network. The performance of the LSTM model is evaluated based on several metrics, and the
predicted prices are compared with the actual stock prices through visualizations.
These values indicate that the LSTM model has a reasonably good performance in predicting stock
prices, with a high R-squared value indicating a good fit to the data.
The LSTM model demonstrates strong potential in predicting stock prices, capturing key trends and
patterns in the data. The performance metrics indicate a reasonable level of accuracy, with room for
further optimization. The visual comparison between actual and predicted prices showcases the model's
ability to follow the general direction of stock price movements.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 25

Dept. of ECE, Dr.AIT, Bangalore-56 Page 26
Dept. of ECE, Dr.AIT, Bangalore-56 Page 27
4.2 APPLICATIONS:

1. Algorithmic Trading:

• High-Frequency Trading (HFT): Utilize predicted stock prices to develop algorithmic

trading strategies that can execute trades at high speed, capitalizing on small price
movements.
• Automated Trading Systems: Implement automated systems that use predictive models to
make buy or sell decisions based on real-time data, reducing human intervention and
increasing efficiency.

2. Portfolio Management:

• Asset Allocation: Use stock price predictions to optimize portfolio allocation by selecting
stocks with higher expected returns and lower risk, improving overall portfolio performance.
• Risk Management: Implement risk management strategies by predicting potential
downturns and adjusting portfolio positions accordingly to minimize losses.

3. Financial Planning and Advisory:

• Investment Advisory: Provide financial advisors and their clients with data-driven insights
for making informed investment decisions, enhancing the quality of investment advice.
• Personal Financial Planning: Help individual investors with long-term financial planning
by predicting stock performance and identifying investment opportunities that align with
their financial goals.

4. Market Analysis:

• Sector Analysis: Predict stock prices within specific sectors to identify trends and
opportunities, aiding investors in making sector-specific investment decisions.
• Market Sentiment Analysis: Combine stock price predictions with sentiment analysis from
news and social media to gauge market sentiment and its potential impact on stock prices.

5. Risk Assessment:

• Credit Risk Modeling: Use stock price predictions to assess the financial health and
creditworthiness of companies, aiding in credit risk management and decision-making for
lenders.
Dept. of ECE, Dr.AIT, Bangalore-56 Page 28
 • Market Volatility Analysis: Predict market volatility and its impact on stock prices to
develop strategies for hedging and managing market risks.

6. Corporate Finance:

• Valuation and M&A: Use predictive models to estimate the future value of companies,
aiding in mergers and acquisitions (M&A) decisions and negotiations.
• Capital Budgeting: Assist companies in making capital budgeting decisions by predicting
the potential impact of investments on stock prices.

7. Regulatory Compliance:

• Market Surveillance: Use predictive models to monitor and detect unusual trading patterns
or market manipulation, aiding regulatory bodies in maintaining market integrity.
• Compliance Monitoring: Assist financial institutions in compliance monitoring by
predicting and analyzing stock price movements to ensure adherence to trading regulations.

8. Research and Development:

• Financial Research: Use predictive models as a tool for academic and industry research to
study market behaviors, test financial theories, and develop new investment strategies.
• Model Innovation: Continuously improve predictive models by incorporating new data
sources, advanced machine learning techniques, and innovative algorithms.

9. Hedge Funds:

• Alpha Generation: Hedge funds can use stock price prediction models to generate alpha by
identifying mispriced stocks and capitalizing on predicted price movements.
• Strategy Development: Develop and backtest new trading strategies based on predictive
models to achieve better risk-adjusted returns.

10. Retail Investment Platforms:

• Robo-Advisors: Integrate predictive models into robo-advisor platforms to offer

personalized investment recommendations and automated portfolio management for retail
investors.
• Investment Apps: Enhance investment apps with predictive analytics features, providing
users with insights and alerts based on predicted stock price movements.
Dept. of ECE, Dr.AIT, Bangalore-56 Page 29
4.3 ADVANTAGES

1. Informed Decision-Making:

• Data-Driven Insights: Stock price prediction models provide data-driven insights, helping
investors make more informed decisions based on historical data and identified patterns.
• Reduced Emotional Bias: Predictive models reduce emotional biases, leading to more
rational and objective investment decisions.

2. Enhanced Investment Strategies:

• Improved Accuracy: Advanced algorithms can capture complex patterns and trends,
resulting in more accurate predictions and better investment strategies.
• Optimal Timing: Predictive models help investors identify optimal entry and exit points,
maximizing returns and minimizing losses.

3. Automation and Efficiency:

• Automated Trading: Predictive models enable automated trading systems that execute
trades swiftly and precisely, increasing efficiency and reducing human error.
• Time-Saving: Automation saves time by quickly processing large volumes of data and
providing actionable insights.

4.Risk Management:

• Risk Mitigation: Predictive models help identify potential risks and market downturns,
allowing proactive measures to mitigate losses and protect investments.
• Volatility Prediction: Accurate predictions of market volatility facilitate better risk
assessment and the development of strategies to manage market fluctuations.

5.Portfolio Optimization:

• Enhanced Performance: Predictive analytics assist in selecting stocks with higher expected
returns and lower risk, leading to optimized and balanced portfolios.
• Effective Diversification: Models can identify correlations between assets, aiding in
effective portfolio diversification to reduce risk.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 30

 6.Competitive Advantage:

• Market Edge: Investors using advanced predictive models gain a competitive edge by
staying ahead of market trends and capitalizing on opportunities.
• Innovation in Trading: Continuous improvements in predictive models foster the
development of innovative trading strategies and techniques.

7. Enhanced Financial Advisory:

• Better Client Advice: Financial advisors can use predictive models to provide accurate and
tailored investment advice, improving client satisfaction.
• Proactive Management: Advisors can manage portfolios proactively by anticipating market
movements and adjusting strategies accordingly.

8. Research and Development:

• Academic Contributions: Predictive models contribute to academic research by providing

tools to test and validate financial theories.
• Financial Innovation: Advancements in predictive analytics drive innovation in financial
technologies and services.

9.Scalability:

• Handling Big Data: Machine learning models efficiently process large volumes of financial
data, making them suitable for various financial institutions.
• Adaptability: Predictive models can be adapted to different markets, asset classes, and
trading environments, offering flexibility and wide applicability.

10.Regulatory and Compliance Benefits:

• Market Surveillance: Predictive analytics assist regulatory bodies in monitoring trading

activities and ensuring market integrity.
• Fraud Detection: Machine learning models help identify fraudulent activities and market
manipulation, enhancing market security.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 31

CHAPTER 5

CONCLUSION AND FUTURE SCOPE

5.1 CONCLUSION

The application of machine learning techniques to stock price prediction has demonstrated
significant potential in enhancing investment strategies and decision-making processes. Through
the use of advanced algorithms such as Long Short-Term Memory (LSTM) networks, predictive
models can capture complex patterns and trends in historical stock data, leading to more accurate
forecasts of future prices.

Key advantages of stock price prediction using machine learning include informed decision-
making, enhanced investment strategies, automation and efficiency, improved risk management,
portfolio optimization, and a competitive edge in the financial markets. These models not only
help in maximizing returns and minimizing losses but also aid in managing market volatility and
mitigating risks.

5.2 FUTURE SCOPE

1. Integration of Additional Data Sources:

• Alternative Data: Incorporating alternative data sources such as social media sentiment,
news articles, macroeconomic indicators, and even satellite imagery can provide richer
context and improve the accuracy of predictions.
• Real-Time Data: Utilizing real-time data streams will enable more dynamic and responsive
predictive models, enhancing their ability to adapt to rapidly changing market conditions.

2. Advanced Machine Learning Techniques:

• Deep Learning Models: Exploring more sophisticated deep learning architectures such as
convolutional neural networks (CNNs) and transformers can uncover new patterns and
relationships in financial data.
• Hybrid Models: Combining different machine learning techniques, such as integrating
LSTM with reinforcement learning or ensemble methods, can leverage the strengths of each
approach for more robust predictions.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 32

 3. Explainability and Interpretability:

• Model Transparency: Developing methods to improve the explainability and

interpretability of machine learning models will help stakeholders understand the factors
driving predictions and build trust in automated systems.
• Feature Importance: Enhancing techniques for assessing feature importance can provide
valuable insights into which variables have the most significant impact on stock price
movements.

4. Personalized Investment Strategies:

• Customized Models: Tailoring predictive models to individual investor profiles, risk

preferences, and investment horizons can result in more personalized and effective
investment strategies.
• Robo-Advisors: Integrating advanced predictive models into robo-advisory platforms can
offer more sophisticated and personalized investment advice to retail investors.

5. Enhanced Risk Management:

• Predictive Risk Models: Developing more accurate models for predicting market volatility
and systemic risks can improve risk management strategies and help in mitigating potential
financial crises.
• Stress Testing: Using predictive analytics for stress testing portfolios under various
hypothetical scenarios can enhance preparedness for adverse market conditions.

6. Global Market Applications:

• Cross-Market Analysis: Extending predictive models to analyze and predict stock prices
across different global markets can provide insights into international market dynamics and
correlations.
• Emerging Markets: Focusing on emerging markets where financial data is becoming more
accessible can open new avenues for research and investment opportunities.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 33

 7. Ethical and Regulatory Considerations:

• Ethical AI: Addressing ethical considerations in the development and deployment of

predictive models, including fairness, transparency, and accountability, will be crucial as
reliance on these models grows.
• Regulatory Compliance: Ensuring that predictive models comply with evolving financial
regulations and standards will be essential to maintain market integrity and protect investors.

8. Quantum Computing:

• Quantum Algorithms: Exploring the potential of quantum computing to solve complex

optimization problems and enhance the processing speed of predictive models could
revolutionize stock price prediction.
• Quantum Machine Learning: Developing quantum machine learning algorithms may
unlock new capabilities in handling vast and complex financial datasets.

9. Collaborative Platforms:

• Crowdsourced Data: Leveraging crowdsourced data and collaborative platforms can

enhance data diversity and model robustness, fostering innovation in predictive analytics.
• Open-Source Models: Promoting open-source development of predictive models can
facilitate knowledge sharing and accelerate advancements in the field.

10. Sustainability and ESG Factors:

• ESG Integration: Incorporating environmental, social, and governance (ESG) factors into
predictive models can align investment strategies with sustainability goals and identify
companies with strong long-term growth potential.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 34

5.3REFERENCE

[1]Usmani M, Adil SH, Raza K, Ali SS. Stock market prediction using machine learning techniques. In
2016 3rd international conference on computer and information sciences (ICCOINS) 2016 Aug 15
(pp.322-327). IEEE.

[2] K. A. Althelaya, E. M. El-Alfy and S. Mohammed, "Evaluation of bidirectional LSTM for short-
and long-term stock market prediction," 2018 9th International Conference on Information and
Communication Systems (ICICS), Irbid, 2018, pp. 151-156.

[3] “Stock price prediction using LSTM, RNN and CNN-sliding window model - IEEE Conference
Publication.” https://ieeexplore.ieee.org/document/8126078 (accessed Dec. 27, 2019).

[4] J. Jagwani, M. Gupta, H. Sachdeva, and A. Singhal, “Stock Price Forecasting Using Data from
Yahoo Finance and Analysing Seasonal and Nonseasonal Trend,” in 2018 Second International
Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India, Jun. 2018, pp.
462–467, doi: 10.1109/ICCONS.2018.8663035.

[5] Donges, N. (2021, August 1). Gradient Descent: An Introduction to 1 of Machine

Learning’sMost Popular Algorithms. Built In. https://builtin.com/data-science/gradient-descent.

[6] Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial
market predictions. European Journal of Operational Research, 270(2), 654-669.

[7]Bao, W., Yue, J., & Rao, Y. (2017). A deep learning framework for financial time series using stacked
autoencoders and long short-term memory. PLoS ONE, 12(7), e0180944.

[8]Patel, J., Shah, S., Thakkar, P., & Kotecha, K. (2015). Predicting stock market index using fusion of
machine learning techniques. Expert Systems with Applications, 42(4), 2162-2172.

[9]Chen, K., Zhou, Y., & Dai, F. (2015). A LSTM-based method for stock returns prediction: A case
study of China stock market. IEEE International Conference on Big Data (Big Data), 2823-2824.

[10]Nelson, D. M., Pereira, A. C., & de Oliveira, R. A. (2017). Stock market's price movement
prediction with LSTM neural networks. International Joint Conference on Neural Networks (IJCNN),
1419-1426.
Dept. of ECE, Dr.AIT, Bangalore-56 Page 35
[11]Brownlee, J. (2018). How to Develop LSTM Models for Time Series Forecasting. Machine
Learning Mastery. Retrieved from https://machinelearningmastery.com/how-to-develop-lstm-models-
for-time-series-forecasting/

[12]Scikit-learn documentation. (n.d.). Metrics and scoring: quantifying the quality of predictions.
Retrieved from https://scikit-learn.org/stable/modules/model_evaluation.html

[13]Yahoo Finance. (n.d.). Historical Stock Prices. Retrieved from https://finance.yahoo.com/

[14]Quandl. (n.d.). Financial and Economic Data. Retrieved from https://www.quandl.com/

TensorFlow. (n.d.). TensorFlow Documentation. Retrieved from https://www.tensorflow.org/Keras.

(n.d.). Keras Documentation. Retrieved from https://keras.io/

[15]Usmani M, Adil SH, Raza K, Ali SS. Stock market prediction using machine learning techniques.
In 2016 3rd international conference on computer and information sciences (ICCOINS) 2016 Aug 15
(pp.322-327). IEEE.

[16] K. A. Althelaya, E. M. El-Alfy and S. Mohammed, "Evaluation of bidirectional LSTM for short-
and long-term stock market prediction," 2018 9th International Conference on Information and
Communication Systems (ICICS), Irbid, 2018, pp. 151-156.

[17] “Stock price prediction using LSTM, RNN and CNN-sliding window model - IEEE Conference
Publication.” https://ieeexplore.ieee.org/document/8126078 (accessed Dec. 27, 2019).

[18] J. Jagwani, M. Gupta, H. Sachdeva, and A. Singhal, “Stock Price Forecasting Using Data from
Yahoo Finance and Analysing Seasonal and Nonseasonal Trend,” in 2018 Second International
Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India, Jun. 2018, pp.
462–467, doi: 10.1109/ICCONS.2018.8663035.

[19] Donges, N. (2021, August 1). Gradient Descent: An Introduction to 1 of Machine Learning’sMost
Popular Algorithms. Built In. https://builtin.com/data-science/gradient-descent.

Dept. of ECE, Dr.AIT, Bangalore-56 Page 36