Pendeteksian Outlier pada Model Regresi Ganda: Studi Kasus Tingkat Penghunian Kamar Hotel di Kendari

Regression is a statistic method to model the correlation between responding variable and free variable. Several things need more attention when making model using regression, one of them is the outlier. Found in the data, outlier will produce less-optimal model. Therefore, outlier detection is needed. This research is objected to detect any outlier as well as identify its influence towards the data used. The outlier detection uses residual value and residual studentized. The data used are about the level of room occupacy in ahotel in Kendari. Two observations are obtained in this research. In addtion, the oulier found causes small R 2 value.