UPM Institutional Repository

Visualization of extracted grammatical role of words using parse tree conversion to improve understanding of English texts


Mirzabeiki, Erfan (2014) Visualization of extracted grammatical role of words using parse tree conversion to improve understanding of English texts. Masters thesis, Universiti Putra Malaysia.


The plethora of English text materials and advancement of digital texts are the factors that have sustained interest in automatic methods for text visualization as an auxiliary tool to improve reading abilities. To better understanding of the English text, firstly, we must be familiar with the English language structure in the text and grammatical role of each word in the sentences and after that, we should be capable of learning, memorizing and understanding the main core of the text. In this study, to identify the grammatical role of the words in text, and improvement of understanding (memorization, recalling, attention extraction, and overall comprehension), an automatic method for grammatical role of the words extraction and visualization is proposed. Our approach identifies the role of each word in the sentences based on generating the dependencies tree and then will extract and visualize each grammatical role of the word in the text based on the user’s requirement. The proposed method for text visualization has the ability of extracting more than 30 grammatical “role of the words” in the text and capable of visualizing the extracted parts using text colour-coding and applying typographic effects (highlighting, bold/italic effects, and font size and font type) on the text. This method aimed at visualization of perceptual organization by uniformly visualize the subjects (role of the words/ role of the words) selected by the users, on the other hand, model of directive attraction and focal factors going to help users to focus and concentrate on the content to be learnt. To validate the effectiveness of extraction and visualization of our work, 40 academic members of Universiti Putra Malaysia (UPM) were asked to give their feedbacks after reading a visualized piece of English text using our proposed method. The result illustrated that compared to other subscales of text visualization; the item which was in higher level of efficiency is the combination of selected visualization items including “font size/colour/bold”, “highlight font size/colour”, and “font style/bold/ colour” format in the text. Therefore, according to the analysis of the experts’ feedback, three groups of visualization: group1 (highlight/size/bold); group 2 (style/colour/size); group 3 (size/colour/bold) performed to design the experiment for this research work to determine the effect of text visualization on user reading abilities. 100 students were the participants of the experiment and they have been assigned into three visualization groups and one control group. The result of Multivariate Analysis Of Variance (MANOVA) revealed that although the text visualization approach for the three groups were more effective on learner's attention extraction, time of skimming, memorization, recalling, and text structure comprehension, the group with combination of (font size/colour/bold style) showed more efficiency compared to other groups. In addition, there is a significant pair wise difference observed between three groups based on reading ability to improve understanding of English text. Moreover, control group scored much lower ratings than the three groups based on understanding abilities (attention extraction, time of skimming, memorization, recalling, and text structure comprehension). Finally, to determine the analysis result of usability factors, all the visualization groups claimed the satisfactory level for each of the three usability scales, which are ease of use, awareness of the structure, and satisfaction of use.

Download File

FSKTM 2014 28IR.pdf

Download (1MB) | Preview

Additional Metadata

Item Type: Thesis (Masters)
Subject: Text processing (Computer science)
Call Number: FSKTM 2014 28
Chairman Supervisor: Assoc. Prof. Lili Nurliyana Abdullah, PhD
Divisions: Faculty of Computer Science and Information Technology
Depositing User: Haridan Mohd Jais
Date Deposited: 08 May 2018 03:20
Last Modified: 08 May 2018 03:20
URI: http://psasir.upm.edu.my/id/eprint/60530
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item