Forecasting Indian inflation using multivariate machine learning

Accurate inflation forecasting is of paramount importance in an inflation targeting monetary policy framework. Advances being made in the field of machine learning are being increasingly adopted in economics to generate inflation forecasting. This study analyses the forecasting performance of multivariate machine learning models in predicting the inflation outcome for India using different algorithms including a linear model (elastic net), a tree-based model (random forest), a feedforward artificial neural network model (multi-layer perceptron) and two recurrent neural network models (Elman network and LSTM). The paper finds that the neural network models outperformed the forecasts of survey of professional forecasters and a random walk model during the periods of study.