Evaluation of lexical cohesion algorithms for Arabic topic segmentation

Abstract EN

The need of having a topic segmentation system for Arabic text is due essentially to improve the functionalities of Arabic Information Retrieval (AIR).

Topic segmentation of texts has been used to improve the accuracy of the subsequent processes such as question answering and information retrieval.

In this paper we present the implementation and the evaluation of two algorithms for Arabic text segmentation which are Text-Tilling and C99.

We compare the quality of the outputs of the two algorithms and we evaluate the relative performance of Text Tiling algorithm with respect to another cohesion based segmenter : C99 algorithm using the classical Recall/Precision evaluation metrics and the recently introduced Reader Judgment method.