Revision history [back]

Hello again,

I have still been trying to understand the code. For this reason, as it was suggested, I am running the code in debug mode. However, without ml 's source code, it seems nearly impossible to understand the code exactly.

I can't understand the info in cascade.xml output file exactly, too. I can understand something of course, but it is not sufficient to comprehend the code totally.

I especially wonder how the weak classifiers are constructed as decision trees and formulated to calculate the error?

Could you explain or at least suggest any sources to understand how the decision trees are formulated to be used in adaboost training?

Thank

Hello again,

I have still been trying to understand the code. For this reason, as it was suggested, I am running the code in debug mode. However, without ml 's source code, it seems nearly impossible to understand the code exactly.

I can't understand the info in cascade.xml output file exactly, too. I can understand something of course, but it is not sufficient to comprehend the code totally.

I especially wonder how the weak classifiers are constructed as decision trees and formulated to calculate the error?

Could you explain or at least suggest any sources to understand how the decision trees are formulated to be used in adaboost training?

ThankThanks.