All clean code presented in *solution.py. There are sometimes too many rows in py files. It should be separated and it will done in my final project. All my steps to understand the assignments are presented in hw.ipynb
In the the following folders you can find:
- lesson 2 is all about metrics used for ranking. I implemented NDCG, Average Precision and pfound
- In lesson 3 I learned all about losses used in ranking like list-wise (Lisnet) or pair-wise (Ranknet). I implemented list-wise Top One Probability net
- In lesson 4 I learned about LambdaMART, YetiRank and learning not a loss but metric directly. I implemented LambdaMART
- In lesson 5 I learned to retrieve documents with KNRM.
- In lesson 6 Currently I'm learning to retrieve closest documents for specific query with navigable small world
⚡ Pending Ranking project including all previous techniques!