Membangun Pythagoras Sebagai Visualisasi Random Forest Untuk Pemodelan Pohon Keputusan

Abstract

Received June 24, 2020 Revised Oct 14, 2020 Accepted Oct 26, 2020 Students observed Pythagoras for using a plane Geometry and 3D Geometry. However, Pythagoras can also be built for decision trees. Our research regarding Instagram Usage Habit with construct Pythagoras for a single decision tree. The study's results obtained are ambiguous attribute values. Therefore, it is continued with research to build Pythagoras for Random Forest. The purpose of the study is to facilitate the tracking of ambiguous data contained in the attributes. The results obtained that the relationship between characteristics of the target class, thus resulting in misclassification. This error caused invalid data; for example, there are three times the separation of data on the same attribute for age's target for a group of 20. However, although there are misclassifications caused by invalid data, based on the Pythagorean construction for Random Forest, the data is more easily traced to errors, which cannot be done by a single decision tree.