An Investigation in Generating Text summary that Adapts to a Specific Reader's Education Level

Faculty Supervisor

Dr. El Sayed Mahmoud

Date of Defense

12-2020

Program Name

Honours Bachelor of Computer Science (Mobile Computing)

School

Applied Computing

Keywords

natural language processing, Flesch Kincaid, machine learning, neural networks, summarization esis

Department

Faculty of Applied Science & Technology (FAST)

Description

Synthesizing a document summary that does not match the reader's education level is challenging and significantly reduces the knowledge uptake.

Abstract

Synthesizing a document summary that does not match the reader's education level is challenging and significantly reduces the knowledge uptake. This work develops an automatic summary generator that adapts the generated summary to a reader's education level. Two summarization systems are developed. The first is a regular summarization system and the second incorporates readability metrics with the first system. The metrics facilitate altering the generated summary to match a specific reader education level. The two metrics: Flesch-Kincaid readability formula and New Dale Chall readability formula were used to avoid biases from each readability formula. The resulting summaries from the two systems are evaluated by a diverse group of readers who match the target education level. The readers complete a short survey that reflects the evaluation for the summaries. The resulting summaries are further evaluated using ROUGE scripts. The result of the surveys are analyzed to evaluate the benefit of incorporating education-level into automatically generated summaries.

Terms of Use

Terms of Use for Works posted in SOURCE.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

This document is currently not available here.

Share Feedback

Share

COinS