|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
NTCIR ProjectTOOLS Home[Japanese]
|Overview| |Evaluation Metrics|Document format conversion|Relevance Assessment Management| Tools
1. Overview of Data formats and Tools (mainly IR)
|
Evaluation tool | run format | doc format | qrel format | |
NTCIR1-6 | trec_eval | TREC format | ntcir_doc format (NTCIR standard document format) |
Early NTCIR format |
NTCIR7- | NTCIREVAL (rev. on July 12, 2023) | res format [sample] |
ntcir_doc format (NTCIR standard document format) |
NTCIR format [sample] |
Results transformation scripts (trec_eval format to ntcir_eval format) |
IR4QA format | |||
TREC format | ||||
PATENT | -trec_eval for PATENT -Precision-oriented evaluation program for NTCIR-5 Passage Retrieval Task |
TREC format | - | - |
*For more details about data formats and tags used, please consult README or README.html included in each Test Collection. |
*The following Test Collections contain the evaluation metrics and scorers:
NTCIR-5 CLQA, NTCIR-6 CLQA, NTCIR-3 QAC, NTCIR-4 QAC, NTCIR-5 QAC
Tool |
Usage |
Task |
Note | ||||||||
ir4qa_eval.tar.gz ir4qa_eval2.tar.gz |
The latest version is included in NTCIREVAL (rev. on July 12, 2023) | 7ACLIA, 8ACLIA, 8CQA | - | ||||||||
NTCIREVAL (rev. on July 12, 2023) | -The latest version- Set of tools for computing various retrieval effectiveness metrics (Binary-relevance metrics such as Average Precision and graded-relevance metrics such as Q-measure and nDCG) |
NTCIR-9 | This is can be used for: NTCIR and TREC Ad Hoc Retrieval, NTCIR-8 CQA, Diversified Search |
||||||||
BOOTS (rev. on May 07, 2016) | BOOTS is a toolkit for conducting pairwise bootstrap tests for a given set of systems, and for computing the discriminative power. | ||||||||||
Discpower (rev. on May 07, 2016) | Discpower is a toolkit for computing the discriminative power of evaluation measures using the randomised Tukey HSD test. | ||||||||||
NTCIRPOOL | Set of simple shell scripts for creating pools for relevance assessments | NTCIR-9 | - | ||||||||
RITE SDK | Java framework for Textual Entailment Recognition system development and evaluation | 9RITE | - |
Tool |
Usage | Lang. |
Task |
||||||||
mai2ntc-r.pl | Convert Mainichi Newspaper Article Data to NTCIR standard document format | J | 7.8ACLIA, 3.4.5.6CLIR, .6CLQA, 8GeoTime, 6MuST, 6OPINION, 7.8MOAT 3.4.5.6QAC, | ||||||||
nyt2ntc.pl |
Convert New York Times Newspaper Article Data to NTCIR standard document format | E | 8GeoTime, 8MOAT | ||||||||
xie2ntc.pl | Convert Xinhua English Newspaper Article Data to NTCIR standard document format | E | 4.5CLIR, 6OPINION, 7.8MOAT | ||||||||
xin2ntc.pl | Convert Xinhua Simplified Chinese Newspaper Article Data to NTCIR standard document format | Cs | 7ACLIA | ||||||||
xin2ntc.1.pl | Convert Xinhua Simplified Chinese Newspaper Article Data to NTCIR standard document format | Cs | 7.8ACLIA, 7.8MOAT | ||||||||
yomi2ntc-clqa.pl | Convert Daily Yomiuri Newspaper Article Data to NTCIR standard document format | E | 5CLIR, 5CLQA, 6OPINION | ||||||||
yomi2ntcir.pl | Convert Yomiuri Newspaper Article Data to NTCIR standard document format | J | 4.5.6CLIR, 5CLQA, 6OPINION, 4QAC |
--- | Test Collection available for research purposes |
--- | Test Collection for the workshop Participants |
System | ||||||||
SEPIA |