Skip to content

Latest commit

 

History

History
69 lines (50 loc) · 3.75 KB

regressions-fire12-en.md

File metadata and controls

69 lines (50 loc) · 3.75 KB
Raw

Anserini Regressions: FIRE 2012 Monolingual English

This page documents BM25 regression experiments for FIRE 2012 ad hoc retrieval (Monolingual English). The document collection can be found in FIRE data page.

The exact configurations for these regressions are stored in this YAML file. Note that this page is automatically generated from this template as part of Anserini's regression pipeline, so do not modify this page directly; modify the template instead.

From one of our Waterloo servers (e.g., orca), the following command will perform the complete regression, end to end:

python src/main/python/run_regression.py --index --verify --search --regression fire12-en

Indexing

Typical indexing command:

target/appassembler/bin/IndexCollection \
  -collection CleanTrecCollection \
  -input /path/to/fire12-en \
  -generator DefaultLuceneDocumentGenerator \
  -index indexes/lucene-index.fire12-en/ \
  -threads 16 -storePositions -storeDocvectors -storeRaw -language en \
  >& logs/log.fire12-en &

The directory /path/to/fire12-en/ should be a directory containing the collection, containing en_BDNews24 and en_TheTelegraph_2001-2010 directories. There should be 392,577 documents in total.

For additional details, see explanation of common indexing options.

Retrieval

Topics and qrels are stored here, which is linked to the Anserini repo as a submodule. They are downloaded from the FIRE data page:

After indexing has completed, you should be able to perform retrieval as follows:

target/appassembler/bin/SearchCollection \
  -index indexes/lucene-index.fire12-en/ \
  -topics tools/topics-and-qrels/topics.fire12en.176-225.txt \
  -topicReader Trec \
  -output runs/run.fire12-en.bm25.topics.fire12en.176-225.txt \
  -bm25 -language en &

Evaluation can be performed using trec_eval:

tools/eval/trec_eval.9.0.4/trec_eval -m map -m P.20 -m ndcg_cut.20 tools/topics-and-qrels/qrels.fire12en.176-225.txt runs/run.fire12-en.bm25.topics.fire12en.176-225.txt

Effectiveness

With the above commands, you should be able to reproduce the following results:

MAP BM25
FIRE 2012 (Monolingual English) 0.3713
P20 BM25
FIRE 2012 (Monolingual English) 0.4970
nDCG@20 BM25
FIRE 2012 (Monolingual English) 0.5420