Olawaleawe/mlinflation

Re-examining Inflation and its drivers in Nigeria: A machine learning approach. # mlinflation ## Re-examining Inflation and its drivers in Nigeria: A machine learning approach. Research on the subject matter of inflation has continued to gain traction in developing countries as over the decades the push and pull forces of inflation have evolved. Most studies examining inflation have adopted traditional econometric and statistical methods such as regression, which assume stringent functional and stochastic processes in fitting data. This study seeks to advance the exploration of inflation drivers in using machine learning algorithm methods capable of relaxing such stringent assumptions, by using the data to discover the best fit. The results from 10 different machine learning models shows that support vector machines, elastic net and robust linear models are the most appropriate algorithms for predicting inflation in Nigeria. Moreover, the bureau de exchange rate, among other important variables, proves to be the most important driver across all models. Therefore, the paper recommends that policy makers should prioritize stabilizing the exchange rate (BCD in particular) in order to achieve price stability.