Daftar Isi: ANALISIS HETEROSKEDASTISITAS PADA MODEL REGRESI LINEAR BERGANDA

Daftar Isi:

Multiple linear regression is relationship function between dependent variable and independent variables. Regression function also involves unknown parameters Y_i= β_0+ β_1 X_1i+β_2 X_2i+⋯+β_k X_ki+u_i Multiple linear regression is heteroscedasti city, if the regression have different variance errors. In contrast, a regression is called homoskedasti city if it has constant variance errors. Regression analysis using heteroscedasti city data will still provide an unbiased estimate for the relationship between the predictor variable and the outcome, but it is inefficient. Biased variance errors lead to biased inference, so results of hypothesis tests are possibly wrong. White test is one of methods to test for the presence of heteroscedasti city. Heteroscedasti city can be removed by a transformation, such as dividing regression be the standard deviation of error term and applying the usual Weighted Least Square(WLS) procedures to transformed regression. This may also improve the approximation to normality. WLS given the estimated parameter is 〖Y_i/σ_i 〗^*= β ̂_0^* 1/σ_i + β ̂_1^* X_1i/σ_i +β ̂_2^* X_2i/σ_i +β ̂_3^* X_3i/σ_i +(e_i )/σ_i