INFLATION FORECASTING FOR PAKISTAN USING MACHINE LEARNING METHODS

Abstract

Economic indicators are vital for any country to make better financial and policy making decisions. Well informed and timely decision are important keeping in view the pace of production, accessibility and supply chains. Economic indicators are dynamic and does not solely rely on financials of the market but also on supply chain, natural calamities and short- and long-term policies by the institutions. We are using statistical models and machine learning models to forecast inflation of Pakistan. A univariate approach is good for getting good forecast, but this is not suitable to capture turning points and identify the causation of inflation. This study involves multivariate approach to forecast consumer price index and to understand the relationship between CPI and other macroeconomic indicators.