Enhancing Forecasting Precision in Indian Industrial Workers' Inflation: A Comparison of Advanced Machine Learning Models

Inflation is a key economic indicator of a country's growth in the contemporary world. This research paper focuses on time series analysis and forecasting of inflation data for Indian industrial workers by employing various machine learning models. This study utilizes the Consumer Price Index (CPI) to measure inflation. It explores and compares the application of Long ShortTerm Memory (LSTM), bidirectional LSTM (BiLSTM), XGBoost, and Random Forest models for predicting future inflation values. The dataset used for this spans from January 1993 to November 2023, sourced from the Labour Bureau of the Government of India. The models are evaluated based on mean squared error (MSE) and their ability to forecast the inflation value for November 2023. The proposed BiLSTM model outperforms the others in terms of MSE and accuracy in forecast values. Although computationally intensive, BiLSTM demonstrates enhanced context understanding and improved learning for long-term dependencies in time series data with a MSE value of 0.9410. The study emphasizes the effectiveness of various forecasting models and the use of data-driven approaches to make decisions based on empirical evidence and inflation trends. It also highlights the potential for incorporating AI into economic analyses for continuous monitoring, effective policy-making, and future advancements.