| Popis: |
A model for automatic text simplification by splitting the longer sentences from a given text has been presented. The paper’s proof-of-concept model demonstrated the following capabilities: (i) the average number of characters per sentence in the output text is reduced to 54% of that of the input text, (ii) 90% of the newly generated sentences are comparable with those generated by a human expert. The Flesch-Kincaid Grade Level (FKGL) of text complexity used in this experiment is 16.5, suitable for graduate students. The FKGL of the processed text is 9.7, suited for tenth grade students. Sentence simplification has various applications in natural language processing, including automatic question generating systems. This work is aimed at simplification so as to enhance accessibility of text and its comprehension by learning disabled children. |
