Comparative Analysis of Time Series Models for Short-term Price Forecasting of Monetary Commodities

Abstract

Time Series Models are used for data that is directly related to time. Unlike regression-based models, these models consider certain time dependencies to provide better forecasting results. Time Series Models are well suited for forecasting stock or monetary commodity prices as these commodities move upward and downward with time. Time Series Models consider seasonal trends and also consider volatility while making predictions on the data. Every monetary commodity has some amount of volatility present in it, especially cryptocurrency which is highly volatile. For these monetary commodities Time Series Models are well suited for future price prediction. Short-term predictions of these monetary commodities can help investors get a general idea about how Gold or Cryptocurrency markets would move in the immediate future. While INR prices in the short term can help determine how it would immediately fare against global currencies. In this research paper, we have compared three different time series models namely ARMA, ARIMA, and Prophet to provide forecasts of three monetary commodities namely INR, Gold, and Cryptocurrency. The results depict a comparison between time series models selected in short-term price prediction of these monetary commodities.