PERBANDINGAN PENERAPAN ESTIMASI METODE REGRESI LINIER MENGGUNAKAN RAPIDMINER DAN MS. EXCEL STUDI KASUS PENJUALAN PADA PUBLIC DATASET

Abstract

Physiological human needs such as clothing, food and shelter can be met through several processes including sales. Sales are activities carried out by companies or individuals to maintain their business in order to develop by getting the desired profit or profit. The tight competition requires companies or individuals to know early the factors that affect the ups and downs of sales volume. Estimated sales volume can be estimated using the Artificial Intelligence branch of science in the form of Data Mining. This research aims to measure the level of ups and downs of sales volume using the multiple liner method. The stages of estimating the rise and fall of sales volume are by determining the dataset taken from public data https://www.kaggle.com/ followed by using data mining methods in the form of estimation and linear regression algorithms applied using ms. excel and rapidminer software each of which produces output Linear regression equation (0.053 * tv + 0.224 * radio + 0. 023 * newspaper) from the results of the process on rapidminer and Equation (3.314 + (0.043 * tv )+ (0.191 * radio )+ (0.002 * neswpaper) from the results of the process on Ms. Excel and its application to the testing data produces an RMSE (root mean squared error) value of 1.744 in the RapidMiner application and 3.277 in the Ms. Excel application, using 160 training data and 40 testing data, or a comparison of 1.744 and 3.277 in the Ms. Excel application. by using 160 training data and 40 testing data, or a ratio of 8:2. Thus, because the RMSE value of the results of Rapidminer is smaller than the results of Ms. Excel calculations, using rapidminer software is more suitable than using Ms. Excel.

Key words: Sales, Regression, Linear, Rapidminer, Data Mining