regression.hc4-v1.0-zh.yaml Maven / Gradle / Ivy

Go to download
---
corpus: hc4-v1.0-zh
corpus_path: collections/multilingual/hc4-v1.0-zh/

index_path: indexes/lucene-index.hc4-v1.0-zh
collection_class: NeuClirCollection
generator_class: DefaultLuceneDocumentGenerator
index_threads: 8
index_options: -storePositions -storeDocvectors -storeRaw -language zh
index_stats:
  documents: 646302
  documents (non-empty): 646302

metrics:
  - metric: MAP
    command: tools/eval/trec_eval.9.0.4/trec_eval
    params: -c -m map
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: nDCG@20
    command: tools/eval/trec_eval.9.0.4/trec_eval
    params: -c -m ndcg_cut.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false
  - metric: J@20
    command: python -m pyserini.eval.trec_eval
    params: -c -m judged.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false
  - metric: Recall@1000
    command: tools/eval/trec_eval.9.0.4/trec_eval
    params: -c -m recall.1000
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false

topic_reader: TsvInt
topic_root: src/main/resources/topics-and-qrels/
qrels_root: src/main/resources/topics-and-qrels/
topics:
  - name: "[HC4 (Chinese): dev-topic title](https://github.com/hltcoe/HC4)"
    id: dev_title
    path: topics.hc4-v1.0-zh.dev.title.tsv
    qrel: qrels.hc4-v1.0-zh.dev.txt
  - name: "[HC4 (Chinese): dev-topic description](https://github.com/hltcoe/HC4)"
    id: dev_description
    path: topics.hc4-v1.0-zh.dev.desc.tsv
    qrel: qrels.hc4-v1.0-zh.dev.txt
  - name: "[HC4 (Chinese): dev-topic description+title](https://github.com/hltcoe/HC4)"
    id: dev_description_title_
    path: topics.hc4-v1.0-zh.dev.desc.title.tsv
    qrel: qrels.hc4-v1.0-zh.dev.txt
  - name: "[HC4 (Chinese): test-topic title](https://github.com/hltcoe/HC4)"
    id: test_title
    path: topics.hc4-v1.0-zh.test.title.tsv
    qrel: qrels.hc4-v1.0-zh.test.txt
  - name: "[HC4 (Chinese): test-topic description](https://github.com/hltcoe/HC4)"
    id: test_description
    path: topics.hc4-v1.0-zh.test.desc.tsv
    qrel: qrels.hc4-v1.0-zh.test.txt
  - name: "[HC4 (Chinese): test-topic description+title](https://github.com/hltcoe/HC4)"
    id: test_description_title
    path: topics.hc4-v1.0-zh.test.desc.title.tsv
    qrel: qrels.hc4-v1.0-zh.test.txt


models:
  - name: bm25-default
    display: BM25 (default)
    params: -bm25 -language zh
    results:
      MAP:
        - 0.2969
        - 0.2030
        - 0.2458
        - 0.1801
        - 0.1455
        - 0.1907
      nDCG@20:
        - 0.3908
        - 0.3023
        - 0.3474
        - 0.2526
        - 0.2048
        - 0.2596
      J@20:
        - 0.4250
        - 0.3800
        - 0.4150
        - 0.3000
        - 0.2470
        - 0.3070
      Recall@1000:
        - 0.7964
        - 0.7255
        - 0.7663
        - 0.6963
        - 0.6358
        - 0.7100
  - name: bm25-default+rm3
    display: +RM3
    params: -bm25 -rm3 -language zh
    results:
      MAP:
        - 0.3202
        - 0.2134
        - 0.2871
        - 0.1625
        - 0.1114
        - 0.1566
      nDCG@20:
        - 0.4262
        - 0.3244
        - 0.3747
        - 0.2149
        - 0.1658
        - 0.2322
      J@20:
        - 0.4000
        - 0.3150
        - 0.3400
        - 0.2790
        - 0.1950
        - 0.2460
      Recall@1000:
        - 0.7589
        - 0.6605
        - 0.7229
        - 0.6814
        - 0.4914
        - 0.5979
  - name: bm25-default+rocchio
    display: +Rocchio
    params: -bm25 -rocchio -language zh
    results:
      MAP:
        - 0.2641
        - 0.2211
        - 0.2641
        - 0.1671
        - 0.1442
        - 0.1929
      nDCG@20:
        - 0.3474
        - 0.2963
        - 0.3502
        - 0.2236
        - 0.1916
        - 0.2644
      J@20:
        - 0.4250
        - 0.3550
        - 0.4250
        - 0.3070
        - 0.2740
        - 0.3350
      Recall@1000:
        - 0.8365
        - 0.7570
        - 0.8241
        - 0.6652
        - 0.6481
        - 0.7074