Python

When i use decision tree as base learner , the model gives R_2 score of 0.9999 (seem overfiting) but when i use linear models even svr with linear kernel, it gives negative value for R2_score im even confuse here