Title :
Does Sequence of Presentation Matter in Reading Comprehension? A Model Based Analysis of Semantic Concept Network Growth during Reading
Author :
Khan, Javed I. ; Hardas, Manas S.
Author_Institution :
Dept. of Comput. Sci., Kent State Univ., Kent, OH, USA
Abstract :
Though not completely natural, reading is one of the most common means of learning for modern humans. In this paper we present an interesting study how the presentation sequence of concepts during reading might impact the semantic network of meaning generated. Text comprehension can be viewed as an incremental and episodic process. Concepts which are recognized are subsequently integrated into the background knowledge represented by a semantic network. We present a reading experiment in which groups of readers are presented same sample texts but in varying order of sentences to simulate different sequences of concept presentation. For comprehensive analysis, we also use a new computational model of the segmentation and integration process and summarize the learning. We then compare the resulting summary networks of the reader groups revealing a number of interesting observations.
Keywords :
computational linguistics; grammars; semantic networks; background knowledge representation; computational model; concept presentation sequence; integration process; model based analysis; reader groups; reading comprehension; reading experiment; segmentation process; semantic concept network growth; sentences; summary networks; text comprehension; Analytical models; Computational modeling; Correlation; Knowledge engineering; Media; Semantics; Text recognition; growth model; reading; semantic analysis; text comprehension;
Conference_Titel :
Semantic Computing (ICSC), 2013 IEEE Seventh International Conference on
Conference_Location :
Irvine, CA
DOI :
10.1109/ICSC.2013.85