Klasifikasi Keatraktifan e-Commerce Indonesia Menggunakan Algoritma Regresi Logistik

Abstract

The rapid development of the internet gives a significant influence in various aspects of life, data mining is the search for value-added processes one of a series of knowledge discovery in databases (KDD). Utilization of data mining can be seen from two aspects, namely commercial and scientific. The collection of test datasets is done by observing the research object on the analysis site that provides vital information and forecast data for the local e-commerce web. The study was conducted by exploration and experimentation, the standard research method used in testing is the cross industry standard process for data mining (CRISP-DM) which consists of 6 phases with the Weka framework 3-8-3. Logistic regression algorithm is a method of data analysis that formulates the response variable with one or more predictor variables. The results of testing the value of e-commerce attractiveness in the confusion matrix, 5 are categorized as having high probability of investment objectives and profiles, 1 are categorized as having good investment probability prospects, 2 are categorized as having good investment probabilities with promising profiles and 1 with promising investment probabilities and profiles not good and 31 in terms of probability and profile is not good. While the results for the Origin class confusion matrix are 31 identified from local Indonesia with confusion 5, 4 from Global with uncertainty 0