This repository contains a Python-based time series analysis project that utilizes ARIMA (AutoRegressive Integrated Moving Average) modeling to forecast number of subscribers . The project aims to demonstrate how to handle, prepare, and forecast time series data using a well-known statistical method
The project follows a systematic approach to time series forecasting:
1.Data Preprocessing: The time series data, which includes a 'Time Period' as the index and 'Subscribers' as a column, is first preprocessed to ensure it is in the correct format for analysis.
2.Training and Testing Split: The dataset is split into training and testing sets based on a temporal cutoff point, ensuring that the model is validated on unseen data.
3.Model Building: An ARIMA model is built using the statsmodels library in Python. The hyperparameters (p,d,q) are determined ACF and PCAF plots.
4.Model Training: The ARIMA model is trained on the training set to understand the underlying patterns in the subscriber count over time.
5.Forecasting: The model makes predictions for the time period covered by the testing set and also make future predictions.
6.Evaluation: The model's predictions are compared against the actual subscriber data in the testing set, and residuals are calculated to evaluate forecasting accuracy.