Dahlan Susilo, Diyah Ruswanti, Supriyanta, Wawan Nugroho

COMPARISON OF PRINCIPAL COMPONENT ANALYSIS AND RANDOM FOREST ALGORITHM FOR PREDICTING HOUSING PRICES

Introduction

Comparison of principal component analysis and random forest algorithm for predicting housing prices. Compare PCA and Random Forest algorithms for accurate housing price prediction in Karanganyar. Discover which method offers optimal results with lower error and faster training.

Abstract

House price predictions are an important thing in the property industry and are useful for buyers in making decisions. Principal Component Analysis (PCA) and Random Forest (RF) methods were used for accuracy analysis in predicting housing prices. Purpose of this research is to measure the accuracy of both methods also to compare RF method optimized with PCA and the one that has not been optimized. The data used is house prices in Karanganyar city based on data scraping results on the rumah123.com site. The analysis reveals that Jaten has the highest number of house sales, and sales of houses with land ownership certificates are also the highest. Of the 10 variables used, land area and buildings have the most influence on selling prices. The model training results show that the RF and PCA methods combination has more optimal value than only using the RF method. The error rate of the PCA method is smaller, averaging 0.0257, making its value more consistent than using only the RF method, which has a larger error value with an average of 0.0332. The model training time using PCA is faster (5005.75) than only using the RF method (6099.25)

Full Text

Comments

You need to be logged in to post a comment.