A Text Analytics-based Approach to Compute Coverage, Readability and Comprehensibility of eBooks

This paper presents our text analytics-based algorithmic formulation for qualitative evaluation of eBooks. Our algorithmic approach computes qualitative aspects of an eBook, particularly its coverage, readability and comprehensibility. Measurement of these qualitative aspects required parsing the eBook text and extracting different information from it such as the concepts described in the eBook. For computing coverage, we have used some standard reference documents in a novel manner. For readability, we have resorted to standard definition and metrics for readability described in some past works. The comprehensibility aspect computation is based on coherent, cohesive and sequentially progressive discussion of learning concepts described in the eBook. The qualitative aspects are quantified in terms of numerical values. The system is tested on a small but diverse dataset of about 30 eBooks. The results are validated by using external information about eBook popularity collected from the Word Wide Web, such as eBook review, and performing sentiment analysis on them. The results obtained give us a fair measure of qualitative account of an eBook on the parameters of coverage, readability and comprehensibility and demonstrates the usefulness of our algorithmic approach.