Windows>> ir-generalized-translation-models>> 返回
项目作者： sebastian-hofstaetter

项目描述 ：
Generalized Translation Models in the Probabilistic Relevance Framework implemented in Lucene & Solr
高级语言： Java
项目地址: git://github.com/sebastian-hofstaetter/ir-generalized-translation-models.git
创建时间： 2017-10-27T11:39:59Z
项目社区：https://github.com/sebastian-hofstaetter/ir-generalized-translation-models
开源协议：Apache License 2.0
下载


Information Retrieval - Generalizing Translation Models

This repository contains a Lucene & Solr implementation of the paper:

Generalizing Translation Models in the Probabilistic Relevance Framework

Rekabsaz, Lupu, Hanbury, Zuccon - CIKM ‘16 (ACM citation)

Implementation

This repository allows to extend Lucene & Solr with a Solr Parser and a Lucene Query. Both are used to extend the search process to include similar terms through an api (the api is not part of this repository)

Extension Structure

Interaction between extension classes of this repository

More information & usage instructions can be found in the Documentation folder

Dependencies

This project depends on the api: https://github.com/sebastian-hofstaetter/embedding_vector_similarity

The api must be running and accessible from the Solr server / Lucene program that uses the extensions from this repository. See the api repository for infos about the similarity computation and possible options (that can be defined in the solrconfig.xml).

Structure

  • Extensions contains the extension classes to be used by Lucene & Solr + self-contained Unit & Integration Tests
  • LuceneEvaluation contains indexing & evaluation code using Lucene for the Extensions (evaluating on TREC-8)
  • Documentation contains explanations of the code / concept and the transformation from the formulas described in the paper to Lucene useable code

The Java projects and their dependencies are managed via maven.