A geometric framework for data fusion in information retrieval

作者：

Highlights：

• A formal geometric framework for data fusion in information retrieval is proposed.

• Score-based data fusion methods such as CombSum and the linear combination method can be well-understood and better used.

• We demonstrate that the conclusions are applicable to both Euclidean distance and ranking-based metrics.

摘要

•A formal geometric framework for data fusion in information retrieval is proposed.•Score-based data fusion methods such as CombSum and the linear combination method can be well-understood and better used.•We demonstrate that the conclusions are applicable to both Euclidean distance and ranking-based metrics.

论文关键词：Database searching,Geometric modeling,Information retrieval,Data fusion

论文评审过程：Received 15 July 2010, Revised 28 August 2013, Accepted 5 January 2015, Available online 12 January 2015.

论文官网地址：https://doi.org/10.1016/j.is.2015.01.001