How to test a machine learning implementation package or a statistical implementation package of NLP algorithms?

Question

I am working on testing several machine learning algorithms, checking to see if they can work as efficiently as described in the documents, and making sure that they can provide great strength to our NLP (Natural Language Processing) statistical platform.

Could you guys show me some test methods for implementing the algorithm? 1) What are the aspects? 2) How? 3) Do I have to follow some basic steps? 4) Do I have to take into account situations with different specifics when using different programming languages? 5) Do I have to understand the algorithm? I mean, does she offer any help if I really know what the algorithm is and how it works?

Basically, we use C or C ++ to implement the algorithm, and our working env is Linux / Unix. Our testing methods focus only on black box testing and function I / O testing. I really want to improve them, but now I do not have a better idea ...

Thanks a lot !! Lol

Answer 1

For many machine learning and statistical classification tasks, the standard metric for measuring quality is Precision and Recall . Most published algorithms will make some complaints about these indicators, or you could implement them and run these tests yourself. This should be a good indicator of the quality you can expect.

, ( Big O). . , , , . , , , , .

, . , , ( , ), , , .