Score Normalization Methods for Relevant Documents Selection for Blind Relevance Feedback in Speech Information Retrieval

Abstract

This paper aims at the automatic selection of the relevant documents for the blind relevance feedback method in speech information retrieval. Usually the relevant documents are selected only by simply determining the first N documents to be relevant. On the contrary, the previous first experiments with the automatic selection of the relevant documents for the blind relevance feedback method has shown the possibilities of the dynamical selection of the relevant documents for each query depending on the content of the retrieved documents instead of just blindly defining the number of the relevant documents to be used in advance. In the first experiments, the World Model Normalization method was used. Based on the promising results, the experiments presented in this paper try to thoroughly examine the possibilities of the application of different score normalization techniques used in the speaker identification task, which was successfully used in the related task of multi-label classification for finding the “correct” topics of a newspaper article in the output of a generative classifier.