Perbandingan Ketepatan Model Logit Dan Probit Dalam Memprediksi Kecenderungan Tingkat Hunian Kamar Usaha Akomodasi Di Bali 2010

Abstract

The aim of this research is to describe how accomodation in Bali looks like and what factors that statistically significant affect accomodation’s room occupancy rate which is the indicator of accomodation productivity with logit and probit equations. Those models then compared to know which one is more precise. The result of this research involving 1.785 accomodations in Bali shows that there is a significant difference among accomodations viewed in different angles, like accomodation type, chain status, and location. From ten variables hipothyzed affecting accomodation’s room occupancy rate, six of themare statistically significant. Those variables are number of workers, number of beds, chain status, association membership, location, and region. Logit and probit models each has 73,39% and 72,94% accuracy. Because of its higher accuracy, logit model is more precise to predict the tendency of accomodation room occupancy rate.