An Unemployment Prediction Rate for Indian Youth Through Time Series Forecasting

The entire process of analysing the trends, seasonality, errors and other features associated with the historical data and deciphering the way to predict future events is known as time series forecasting. Here, the previous data would be present with the respective time. The format of time could be anything annual, weekly, daily, etc. The objective of this work is to predict the youth unemployment rate in India for five years from 2019 to 2023. The data on the youth unemployment rate in India from 1991 till the end of 2018 is taken. In this work, the frequency of the time is annual. After the analysis of historical data available, various models are fitted onto this time series individually, for the purpose of predicting the future unemployment rate. We have compared several exponential smoothing models and ARIMA model onto the time series data and have come across that the ARIMA model performs best among all the models on the basis of forecasting accuracy via time series cross-validation.