Neural networks are not "off-the-shelf" algorithms in the way that random forest or logistic regression are. There are two features of neural networks that make verification even more important than for other types of machine learning or statistical models. You have to check that your code is free of bugs before you can tune network performance! Otherwise, you might as well be re-arranging deck chairs on the RMS Titanic. Writing good unit tests is a key piece of becoming a good statistician/data scientist/machine learning expert/neural network practitioner. This can be done by comparing the segment output to what you know to be the correct answer. The best method I've ever found for verifying correctness is to break your code into small segments, and verify that each segment works. For programmers (or at least data scientists) the expression could be re-phrased as "All coding is debugging."Īny time you're writing code, you need to verify that it works as intended. There's a saying among writers that "All writing is re-writing" - that is, the greater part of writing is revising.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |