Time Series Forecasting of Stock Prices

This project focuses on employing the Auto-Regressive Integrated Moving Average (ARIMA) model to forecast stock prices. It utilizes historical stock data of Bajaj Finance (BAJFINANCE) obtained from a CSV file. The key objective is to develop a predictive model capable of accurately forecasting stock prices, aiding investors and analysts in making informed decisions.

● Visualization: Matplotlib and Seaborn libraries are employed to visualize the actual stock prices alongside the ARIMA model's forecasts.

● Model Persistence: The model is serialized using the pickle module, allowing for easy storage and retrieval. A file named 'stock_arima.pickle' contains the serialized model, ensuring reusability without the need for retraining.

● Model Selection and Training: Used ARIMA model for its effectiveness in modelling time-series data and pmdarima library’s ‘autoarima’ function is utilized to automatically select the optimal parameters (p, d, q) for model through exhaustive search.