Skip to content

The project consists in the development of some Rank Fusion algorithms and the subsequent evaluation of these algorithms. A TREC dataset that provides documents, topic and relevance judgements is used as collection and Information Retrieval Evaluation mea- sures are applied to evaluate the performances. Rank Fusion algorithms combine a set of or…

Notifications You must be signed in to change notification settings

LcRss/Information_Retrieval_Project_ProbFuse

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 

Repository files navigation

README

INFO

This code was implemented during an Information Retrieval course by:

Luca Rossi; Davide Storato; Roberto Tarantini; Andrea Ziggiotto.

PDF Report: https://drive.google.com/file/d/1595CxbW8jsgRC4e0pkH-A2aSmLCLxNJH/view?usp=sharing

Slide: https://drive.google.com/file/d/1UDXI_J3Wu64rMMPQDPS22IC6L0FiGJUo/view?usp=sharing

COMB METHODS

The class ComputeRankFusion is used to execute the base methods, the parameters PathInput and PathOutput must be set in this way:
-PathInput: is the directory where are saved all the original runs;
-PathOutput: is the directory where the fusion run will be saved.
This class computes all Comb methods at once, the output is compound of 6 different txt files in TREC format.

PROBFUSE

The class ComputeProbFuse is used to execute ProbFuse, the parameters pathRun, pathQrels, pathOutput must be set in this way:
-pathRun: is the directory where are saved all the original run;
-pathQrels: is the position where is saved the file of the qrels;
-pathOutput: is the directory where will be saved the txt file of the final run.
To executes this software, it must be set also the dimension of the training size and the number of segments. These value can be set in the call of the ProbFuse class constructor.

About

The project consists in the development of some Rank Fusion algorithms and the subsequent evaluation of these algorithms. A TREC dataset that provides documents, topic and relevance judgements is used as collection and Information Retrieval Evaluation mea- sures are applied to evaluate the performances. Rank Fusion algorithms combine a set of or…

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages