3

I have seen some classes in my University labeled as an "Inference class" and others as "Machine learning" classes, but I was not sure if appreciated the core difference between these two labelings?

I normally think of machine learning as the study of systems/algorithms that are able tp extract information from data. It uses techiniques from a broad list subjects, like statistics, computer science, optimization, etc. However, I was not sure how statistical inference differed from Machine Learning.

Thanks for the help in advance! :)

Franck Dernoncourt
  • 42,093
  • 30
  • 155
  • 271
Charlie Parker
  • 5,836
  • 11
  • 57
  • 113
  • 3
    I think the second answer to this question http://stats.stackexchange.com/questions/6/the-two-cultures-statistics-vs-machine-learning also answers your question. – david25272 Feb 03 '14 at 04:25
  • Thats interesting, the difference seems a little subtle, because machine learning seems a bit more focused on having a good prediction function, while inference focuses more on how to learn the process of data generation. But if you can generate, you should be able to predict well!! Or at least that claim seems to make sense at a first glance. i.e. one cares on getting good predictions while the other cares more on getting good *understanding* of how predictions (i.e. the process) are made. That's my understanding so far. – Charlie Parker Feb 03 '14 at 04:46
  • 2
    @Pinocchio That's probably not a bad assessment of the difference (stats doesn't avoid black boxes, but much of it tends to look for non-black-box approaches; ML jumps right into the black box -- neither is doing it right or wrong, they're emphasizing different problems). Of course, this is oversimplifying hugely - statistical inference isn't *only* concerned with prediction for example, but with a variety of other tasks (point and interval estimation, hypothesis testing, ...) – Glen_b Feb 03 '14 at 05:41
  • My question has to do specifically with inference, not with statistics in the more general sense. I am not sure if it just reduces to the question that was posted at the top, but I feel that the question does not address explicitly what I was looking for because I read that before posting this question. Maybe it's just my own ignorance on the topic, but I still want to know the difference of what the word inference means and how exactly its different from machine learning. – Charlie Parker Feb 03 '14 at 17:11
  • I think this question is actually more related than the post marked as "duplicate": https://stats.stackexchange.com/questions/130867/inference-vs-estimation?rq=1 – Charlie Parker May 27 '17 at 04:12

0 Answers0