The web-based estimation of motorcycles sales using linear regression method

Abstract

Abstract The purpose of this research is to use the linear regression method to predict motorcycle sales results, the variable used is the period as an independent variable (X) and sales as the dependent variable (Y). The data used in the calculation of linear regression is the last four years data, from January 2014 to December 2019. The implementation of the motorcycle sales forecasting system is to predict sales in the coming months. To find out the level of accuracy of the prediction error calculation is needed so that it is known how many error levels are obtained. Calculation of forecasting errors using Mean Absolute Deviation (MAD) and Mean Absolute Percentage Error (MAPE). The results of this study are web based motorcycle sales prediction systems using linear regression method. From this system, motorcycle sales forecasting is obtained the following month. In January 2015 with forecasting results of 12.63. To find out how accurate the forecasting level is, the error calculation result using Mean Absolute Deviation (MAD) is 3.40 and Mean Absolute Percentage Error (MAPE) is 44.33%. The results show that the error rate is small and the forecasting results are close to accurate.