Forecasting Consumer Price Index (CPI) Using Deep Learning and Hybrid Ensemble Technique

In today’s day and age, economic crises are all over the world due to high inflation. Inflation is a rise in price of the goods and services produced in a country. As a result of rising prices, a given amount of money can now buy fewer goods and services. The general public’s cost of living is affected by this loss of purchasing power, which ultimately slows economic growth. Thus, it has a negative impact on the purchasing power of the people. Various sorts of baskets of commodities are generated and tracked as price indices to calculate inflation or deflation, depending on the chosen set of goods and services used. One type of price index proposed in this project is Consumer Price Index (CPI), which looks at the weighted average of costs for a variety of products and services like transportation, food, and healthcare. This paper proposes different deep learning time series models such as LSTM, BiLSTM and hybrid ensemble learning to forecast the Indian consumer price index (CPI). These two single RNN models (LSTMs and BiLSTMs) are compared with the hybrid ensemble learning model to see which gives better forecasting results for the consumer price index.