Hi Asmara,

The margin is defined as the difference between the probability predicted for the actual class and the highest probability predicted for the other classes. One hypothesis as to the good performance of boosting algorithms is that they increase the margins on the training data and this gives better performance on test data.

HTH.

Samer, PhD.

Thanks Sir.

when the predicted margin is 1 whats does its mean ?

and when the predicted margin is in negative number what does it mean?

A margin of 1 means that the correct class is predicted with 100% confidence (very good), a margin of -1 means that an incorrect class is predicted with 100% confidence (very bad).

Cheers,

Samer

Sir 0ne more question

0.238095

-0.64539

-0.64539

-0.92179

0.382199

0.908046

0.744681

when the predicted margin is like these value some are negative and some are positive but they are neither 1 not -1 so what does these value represent ?

From my reply you can realize that the results are in the range [-1, 1]. The larger the margin, the more confident the classifier is in predicting the true class. For instance, a negative margin implies that the corresponding instance is misclassified.

Cheers,

Samer