Analisis Regresi Pada Pelanggaran Asumsi Klasik Pada Regresi Linear

Abstract

Linear regression model analysis requires the fulfillment of various assumptions so that the model can be used as a good predictive tool. However, it is not uncommon for researchers to face problems in their models. Several violations of classic assumptions that often occur in OLS estimation, namely violations of the homoscedasticity assumption and the occurrence of autocorrelation. The method used in this research is descriptive qualitative with data collection techniques using library research. This research aims to determine whether there are violations of classical assumptions in a linear regression model, how to detect them and how to overcome them. And we are expected to know the impact of these deviations or violations. The results of this research found that the classical assumption tests that are often used are the multicollinearity test, heteroscedasticity test, normality test, autocorrelation test and linearity test. There are no definite provisions regarding the order in which tests must be fulfilled first. Analysis can be carried out depending on the existing data. For example, an analysis of all classical assumption tests is carried out, then we see which ones do not meet the requirements. Then improvements are made to the test, and after meeting the requirements, another test is carried out.