Accuracy at the time of training

Hi,

I am doing some training of data using SVM. However at the end of taining, it shows me some accuracy. What is meant by accuracy at the time of training? Does it mean that the kernel also perform some sort of prediction on the training data?

Please guide me.

Zulfi.

Answer 1

[Walter Roberson]

The training process aggregates the data as best it can, knowing what the proper result will be. However, the training process is seldom perfect. The training process continually calculates accuracy as it goes, trying to maximize accuracy. Eventually it gives up trying to make improvements, and reports the accuracy it has obtained.

[Zulfiqar Khan]

Hi,

Thanks for your response. In artificial neural network (ANN) the training process updates the weights. So what the training process does in SVM to improve the accuracy?

Zulfi.