Application of Forecasting Model to Study the Population Growth of India | Chapter 4 | Mathematics and Computer Science: Research Updates Vol. 4

This study aims to compare the forecasting accuracy of three-time series models—Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing State Space Model (ETS), and State-Space Model with Kalman Filtering—for predicting India’s population trends from 2021 to 2091.

Study Design: A comparative analysis of time series forecasting models based on historical population data.

Methodology: The study employs statistical and machine learning forecasting models: ARIMA, ETS, and Kalman Filtering. These models were applied to obtain forecasts and compared based on their errors and alignment with currently available data.

Results: The ARIMA model estimates a decline in India's population after 2051, while the ETS and Kalman Filtering models suggest continuous growth until 2091. The ARIMA model shows the lowest error rate (MAE: 14.63, RMSE: 24.24, MAPE: 3.36%), providing better short-term accuracy. The ETS model offers a more reliable long-term projection, though with slightly higher errors. The Kalman Filtering model presents the highest error values (MAE: 42, RMSE: 54.06, MAPE: 12.67%), reflecting greater uncertainty in its estimates.

Conclusion: Among the statistical models, ARIMA delivers the most accurate short-term forecasts, while ETS and Kalman Filtering are more suitable for long-term projections. In the machine learning forecast, XGBoost was found to be the most accurate for long-term population forecasting. The study highlights the strengths and limitations of each model and underscores the importance of selecting an appropriate forecasting method based on the required time horizon and accuracy needs.

 

Author (s) Details

 

Abhishek Pandey
Department of Computer Science and Engineering, Institute of Advanced Research, The University for Innovation, Gandhinagar, Gujarat. Pin :382426, India.

 

Sanjay Sonar
Department of Computer Science and Engineering, Institute of Advanced Research, The University for Innovation, Gandhinagar, Gujarat. Pin :382426, India.

 

Please see the book here:- https://doi.org/10.9734/bpi/mcsru/v4/4778