Skip to content

ruisizhang123/Pseudo-Relevance-Feedback

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Feature

Ranking documents of a query using BM25 Score in Document Ranking Phase and Rocchio Algorithm in Query Expansion Phase.

Usage

  1. Create a folder name data and put query txt and doc txt in ./data folder

  2. Run EE448.ipynb to visual output

  3. The output ranked documents is in ./data/bm25_score.txt

Dependencies

Python >= 3.0

Dataset Overview

You can get dataset here or use you own data.

  1. ./data/query.txt: query_id \t query_text

  2. ./data/doc.txt: document_id \t document_text

To Get Better Ranking Results...

  1. Set expansion words in util.py/findNewQuery/loopRange to different value. If the documents is short, set loopRange to a smaller value.

  2. Set k2 in score.py/bm25 to larger value.

  3. Set GAMMA to 0.15 or 0 to enable positive feedback and negative feedback

  4. You may try different Score Function like TF-IDF to rank documents in score.py

Reference

https://github.com/preethiA30/Search-Engine

Releases

No releases published

Packages

No packages published