Future Live Stock Prediction using Large Language Models in Python.

Stock market prediction has always been a challenging task for investors and traders. However, recent advancements in natural language processing and deep learning have opened up new possibilities. In this article, we will explore how we can leverage large language models and Yahoo Finance data to make future stock price predictions using Python.

Large Language Models Article (LLM)

Table of Contents:

1. Understanding the Problem
2. Gathering Historical Stock Data
3. Preprocessing the Data
4. Using a Large Language Model
5. Generating Future Stock Price Predictions
6. Conclusion

Section 1: Understanding the Problem To begin, let’s understand the problem we aim to solve. We want to predict future stock prices based on historical data. While this task is inherently uncertain, language models can help us analyze patterns and generate speculative predictions.

Section 2: Gathering Historical Stock Data We start by using the “yfinance” library in Python to fetch historical stock data from Yahoo Finance. By specifying the desired ticker symbol and date range, we can retrieve a dataset containing stock prices and corresponding dates.

Section 3: Preprocessing the Data Next, we preprocess the fetched data to extract the necessary information. We focus on the closing prices and dates, which serve as our key features for prediction. We organize the data into a suitable format for further analysis.

Section 4: Using a Large Language Model In this section, we introduce the concept of large language models. These models, such as GPT-3, have been trained on vast amounts of text data and can generate coherent and contextually relevant responses. We leverage the power of large language models to analyze historical stock data and generate predictions.

Section 5: Generating Future Stock Price Predictions To generate future stock price predictions, we use the OpenAI API and a suitable language model (e.g., GPT-3.5 Turbo). We pass the preprocessed historical data as input and specify the desired number of predictions. The language model then generates speculative predictions based on the patterns and trends in the provided data.

Section 6: Conclusion In this article, we explored how we can use large language models in Python to make future stock price predictions based on historical data from Yahoo Finance. While these predictions are speculative and should not be considered financial advice, they offer an interesting approach to analyzing and interpreting stock market trends. Remember that investing in the stock market involves risks, and it is crucial to consult with financial professionals before making any investment decisions.

Practice Python Code example:

import yfinance as yf
import openai
import matplotlib.pyplot as plt

ticker = "^NSEI"  
start_date = "2023-01-01"
end_date = "2023-06-08"
data = yf.download(ticker, start=start_date, end=end_date)

data["Date"] = data.index
data.reset_index(drop=True, inplace=True)

subset_data = data.sample(10) 
subset_prices = subset_data["Close"].tolist()
subset_dates = subset_data["Date"].dt.strftime("%Y-%m-%d").tolist()

text = "\n".join([f"{date}: {price}" for date, price in zip(subset_dates, subset_prices)])

openai.api_key = "YOUR_API_KEY"  

num_predictions = 5  
response = openai.Completion.create(
    engine="text-davinci-003",  
    prompt=text,
    max_tokens=50,
    n=num_predictions,
    stop=None,
)
predicted_prices = []

for i, choice in enumerate(response.choices):
    generated_text = choice["text"].strip()
    prediction = generated_text.split(":")[-1].strip()
    # print(f"Prediction {i+1}: {prediction}")
    predicted_prices.append(prediction)
    print(predicted_prices)


last_date = dates[-1]
predicted_dates = np.arange(np.datetime64(last_date), np.datetime64(last_date) + np.timedelta64(num_predictions + 1, 'D'), np.timedelta64(1, 'D'))
predicted_dates = predicted_dates[1:]  
predicted_dates = [str(date) for date in predicted_dates]  

plt.figure(figsize=(12, 6))
plt.plot(dates, prices, label="Historical Prices")
plt.plot(dates[-1], prices[-1], "ro", label="Last Historical Price")
plt.plot(predicted_dates, predicted_prices, "g^", label="Predicted Prices")
plt.xlabel("Date")
plt.ylabel("Stock Price")
plt.title(f"{ticker} - Historical and Predicted Stock Prices")
plt.legend()
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

Result of Price Prediction

Visit my blog: https://medium.com/@sahajgodhani777

Visit my official website: https://sahajgodhani.in/