DW is generally used to test first-order autocorrelation

LM can be used to test first-order and higher-order autocorrelation, and in general in eviews, we only test to the second order, you can check the DW statistic in the regression results with lag two, if it is very close to two, then it means that autocorrelation is eliminated. Eliminate the autocorrelation, at the same time, you also need to check the chi-square distribution percentile distribution table, to see if LM = (T * R-square) is greater than the critical value, if it is that there is a second-order autocorrelation

Then use the generalized difference method to eliminate the autocorrelation on it