backgroundlinking19.yaml

---
corpus: wapo.v2
corpus_path: collections/newswire/WashingtonPost.v2/data/

index_path: indexes/lucene-index.wapo.v2/
collection_class: WashingtonPostCollection
generator_class: WashingtonPostGenerator
index_threads: 1
index_options: -storePositions -storeDocvectors -storeRaw
index_stats:
  documents: 595031
  documents (non-empty): 595030
  total terms: 318219870

metrics:
  - metric: MAP
    command: target/appassembler/bin/trec_eval
    params: -c -M1000 -m map
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: nDCG@5
    command: target/appassembler/bin/trec_eval
    params: -c -M1000 -m ndcg_cut.5
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true

topic_reader: BackgroundLinking
topics:
  - name: "[TREC 2019 Topics](https://github.com/castorini/anserini-tools/tree/master/topics-and-qrels/topics.backgroundlinking19.txt)"
    id: bglink19
    path: topics.backgroundlinking19.txt
    qrel: qrels.backgroundlinking19.txt

models:
  - name: bm25
    display: BM25
    params: -backgroundLinking -backgroundLinking.k 100 -bm25 -hits 100
    results:
      MAP:
        - 0.3029
      nDCG@5:
        - 0.4785
  - name: bm25+rm3
    display: +RM3
    params: -backgroundLinking -backgroundLinking.k 100 -bm25 -rm3 -hits 100
    results:
      MAP:
        - 0.3787
      nDCG@5:
        - 0.5200
  - name: bm25+rm3+df
    display: +RM3+DF
    params: -backgroundLinking -backgroundLinking.dateFilter -backgroundLinking.k 100 -bm25 -rm3 -hits 100
    results:
      MAP:
        - 0.3160
      nDCG@5:
        - 0.5018