Learning English Grapheme Segmentation Using the Iterated Version Space Algorithm

Abstract. Our unique approach for learning English grapheme segmentation (LE-GS) rules using the Iterated Version Space Algorithm (IVSA) is presented. After de ning the problem and our representation for the instances and hypotheses, we illustrate the LE-GS approach by tracing a speci c example. Experimental results, based on a ten-fold testing methodology, are given to show the performance of the LE-GS learning system.

• Type: Conference
• Year: 1999
• Where: ISMIS
• Authors: Jianna Jian Zhang, Howard J. Hamilton, Nick Cercone
• Keywords: Artificial Intelligence, English Grapheme Segmentation, ISMIS 1999, Iterated Version Space, Ten-fold Testing Methodology