Is my model difference between Training accuracy and cross validation accuracy considered as overfit?

Question

Is my model difference between Training accuracy and cross validation accuracy considered as overfit?

Amben Uchiha

2022年2月1日 04:28

So I used weka to determine my training accuracy and cross-validation accuracy. It has showed me that my training accuracy is 84.9167 % and my Cross validation accuracy is 83.9167 %

I also tried to use sklearn to determine my training and cross validation accuracy and gave me the following: 83.5% on training and 82.67% on cross validation accuracy.

Is the difference between training accuracy and cross validation accuracy enough to consider my model overfit?

Topic weka scikit-learn

Category Data Science

Ashwiniku918 · Accepted Answer · 2022年2月1日 04:28

Absolutely no. Difference between training and cross validation is expected in most scenarios as training use almost all the data and cross validation misses on some sample.

Overfitting happens when the difference between training and testing data is too high. In your case neither the diff is too high nor testing data is used.

Is my model difference between Training accuracy and cross validation accuracy considered as overfit?

About