Let me take an explicit example.(the answer should be universal for any algorithm) I want to group cities with similar average income together I want to use dbscan to do this.(R/python/javascript/spark)

But suppose I don't know what is dbscan The next thing I need to do is either: 1 read a book about it and get deep understanding then apply it. 2 just read how to use the method, apply it without any understanding 3 Read the wiki or any short explanation then apply it.

1 is very slow, it may takes you weeks to read a book and sometimes you don't have so much time.

2,3 is fast but since you didn't quite understand the full picture, you may find some bug in future when you apply, or may be even worse since you don't fully understand the algorithm you can't even tell if there is a bug.

So my question is in this situation, what should you do to make you have enough info to apply a algorithm without read a 800 pages book about it

I would hire someone or otherwise admit that this wasn't possible.

That is, when presented with an algorithm (or anything else, really) that you are unfamiliar with you either a) Have enough knowledge to be able to learn about it quickly enough to suit your purposes or b) Not a). In the case of not a) the thing to do is admit that you need more time or else hire a specialist.

Your Answer

 
discard

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.