Section 1.1: Setting Up the Environment

To kick off our project, we need to collect genuine financial data. We will use the yfinance library to acquire historical price information for a varied selection of securities available on Yahoo Finance. Begin by installing the library with the following command:

pip install yfinance

Now, let’s import the essential libraries in Python, including numpy for numerical tasks and yfinance for retrieving financial data. Additionally, we’ll import matplotlib for visualizing the data. Include these imports in your Python script:

import numpy as np

import yfinance as yf

import matplotlib.pyplot as plt

Section 1.2: Acquiring Historical Price Data

Next, we will download historical price data for several securities until the end of February 2024. To ensure a comprehensive market representation, we will select Tesla (TSLA), Amazon (AMZN), and Bitcoin (BTC-USD) for our analysis.

# Download historical price data for Tesla, Amazon, and Bitcoin

tesla_data = yf.download('TSLA', start='2020-01-01', end='2024-02-29')

amazon_data = yf.download('AMZN', start='2020-01-01', end='2024-02-29')

bitcoin_data = yf.download('BTC-USD', start='2020-01-01', end='2024-02-29')

With our data in hand, let’s visualize the closing prices of these securities over time. We will create a plot using matplotlib to gain insights into their price trajectories.

# Plotting the closing prices of Tesla, Amazon, and Bitcoin

plt.figure(figsize=(14, 7))

plt.plot(tesla_data['Close'], label='Tesla')

plt.plot(amazon_data['Close'], label='Amazon')

plt.plot(bitcoin_data['Close'], label='Bitcoin')

plt.title('Historical Closing Prices of Securities')

plt.xlabel('Date')

plt.ylabel('Price (USD)')

plt.legend()

plt.grid(True)

Section 2.1: Model Construction

First, we need to import the necessary components from Keras to build our model. We will utilize the Sequential model to arrange layers sequentially.

from keras.models import Sequential

from keras.layers import LSTM, Dense

Next, we must preprocess our data prior to inputting it into the model. Normalizing the closing prices is essential to ensure they are comparable. Here’s a function to accomplish this:

def normalize_data(data):

return (data - np.min(data)) / (np.max(data) - np.min(data))

With normalized data, we can now divide it into training and testing sets, using Tesla’s historical price data for this example.

# Normalize the closing prices of Tesla

tesla_close = tesla_data['Close'].values.reshape(-1, 1)

tesla_close_norm = normalize_data(tesla_close)

# Split the data into training and testing sets

train_size = int(len(tesla_close_norm) * 0.8)

train_data = tesla_close_norm[:train_size]

test_data = tesla_close_norm[train_size:]

Section 2.2: Sequence Creation for Predictions

To enable the model to make predictions, we need to create sequences from the historical prices. Below is a function to generate these sequences based on a defined length.

def create_sequences(data, seq_length):

X, y = [], []

for i in range(len(data) - seq_length):

X.append(data[i:i+seq_length])

y.append(data[i+seq_length])

return np.array(X), np.array(y)

With sequences established, we will build our RNN model using Keras. We will create a simple LSTM model comprising one LSTM layer and a dense output layer.

# Define the LSTM model

model = Sequential()

model.add(LSTM(64, input_shape=(5, 1)))

model.add(Dense(1))

# Compile the model

model.compile(optimizer='adam', loss='mean_squared_error')