Prediction The Price of National Groceries Using Average Based Fuzzy Time Series With Song - Chissom and Markov Chain Approach

Abstract

Groceries are strategic commodities that have an important role in economic, social, and even political aspects in various countries including Indonesia. The groceries affect the livelihood of the people with the scale of the fulfilment of high needs as well as factors supporting the welfare of the community. The classical problem in the fulfilment of grocery is the fluctuation of the prices of groceries. The increase in the prices of groceries commodities becomes a major factor in inflation. To overcome these problems, one of the efforts made by the government is to stabilize the price policy of grocery so that farmers as producers get profitable results and the community as consumers can afford to buy groceries at affordable prices. To accommodate the afford it is needed a forecasting step to predict the prices of groceries. This study aims to predict the prices of national groceries using the Average Based Fuzzy Time Series method with Song - Chissom and Markov Chain approach. The data used are prices of groceries weekly period from 2015 - 2017. Data is divided into two phases: training and testing dataset with the ratio of 90: 10. Based on MAPE value and feasibility test, it can be concluded that Average Based Fuzzy Time Series with Markov Chain approach shew better than Song - Chissom approach for prediction the prices of national groceries.