Effects of the College Major on Assessments of Arabic Text

Summaries

Bassam Hammo¹, Martha Evens², and Hani Abu-Salem³

¹Department of Computer Information Systems, the University of Jordan, Jordan

²Department of Computer Science, Illinois Institute of Technology, USA

³Department of Mathematics and Computer Science, University of South Carolina-Aiken, USA

Abstract: We set out to discover whether or not the summaries produced by our Arabic text summarization software were potentially useful to a wide range of people. 1200 students at the University of Jordan were each given a copy of a newspaper article and a system-generated summary and asked to classify the summary as Rejected (R), Not-Related (N), Satisfactory (S), Good (G) or Accepted (A). 76.92% of the summaries were judged to be good or accepted and 92.34% were judged to be satisfactory, good or accepted. These students came from four different majors: 300 from Arabic studies, 300 from humanities, 300 from Information Technology (IT) and 300 from a one-year program designed to help K-12 teachers to learn how to use computers effectively in the classroom. To our surprise, students from these four different majors differed significantly in their assessments; the teachers rated the summaries significantly more favourably; the IT students rated them significantly lower than did the students in Arabic and the humanities.

Keywords: Arabic natural language processing, arabic text summarization, extraction, software testing, evaluation.

Received June 18, 2014; accepted December 16, 2014