TAG LINES EXTRACTION AND RANKING IN TEXT ANNOTATION

Abstract

The article deals with comparative analysis of two approaches for tag lines ranking in text annotation task. The first approach is based on a weight evaluation of extracted phrases using TextRank algorithm, the second one is based on tf-idf estimation. Experiments were carried out on the base of INSPEC dataset collection. Experiment descriptions and comparative results are given. It was shown experimentally that tf-idf approach gives better results than the one based on TextRank.

Keywords

• annotation
• tag lines extraction and ranking
• annotation quality evaluation