An attention encoder-decoder RNN model with teacher forcing for predicting consumer price index

The aftermath of the COVID-19 pandemic has led to a global surge in inflation rates across the world, eroding consumer purchasing power and sparking concerns of potential recession. A critical gauge of inflation is the Consumer Price Index (CPI), which measures the average price change of different baskets of goods and services over time. Accurate forecasting of CPI is crucial for policymakers to strategically manage inflation as well as take well informed decisions around fiscal and monetary policy. In this paper, we propose a novel deep learning approach for forecasting the Canadian consumer price index (CPI) utilizing a recurrent neural network encoder-decoder attention model that incorporates the teacher-forcing technique. To the best of our knowledge, this paper is the first to examine the value of encoder-decoder attention models for inflation prediction, shedding light on the significance of advanced deep learning models in macroeconomic predictions. Our results demonstrate the efficacy of this model in accurately forecasting CPI due to its capacity to capture long-term dependencies in inflationary factors. Our model outperforms well-established benchmarks from traditional statistical learning, machine learning, and deep learning methods, including ARIMA, Lasso regressor, support vector regressor, random forest, and artificial neural networks, consistently minimizing forecasting errors across all CPI indicators. The proposed model achieves 45% improvement in RMSE score compared to the best performing machine learning benchmark model. Forecasting results suggest an increase in CPI across all categories, with the energy, food, and gasoline exhibiting the sharpest increase.