Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
I am using GridSearchCV to find the best combination of parameters for SVM. However, the parameters chosen by GridSeasrchCV do not seem to be the best ones. I tried some parameters randomly and they are working much better. I tried both combinations in the test set. Here is the picture:
The second confusion matrix + classification report is the random but improved one. Why isn't GridSearchCV selecting this combination?
99% Accuracy is over fitted model and its not the accurate model. There will always be loss associated. secondly you sample size is too small for reaching the global minima. you should also check your ROC-AUC curve for the model.
