Sungjae Han

A Personalized Assessment System Based on Item Response Theory

Computerized adaptive testing (CAT) is a method of administering tests that adapts to the examinee’s ability level. Previous research has focused on estimating the examinee’s ability accurately and on providing adequate feedback upon analyzing the examinee’s ability. However, in order for students to use the feedback, they must find courses or learning materials themselves. It is difficult to make customized learning available continuously. Therefore, we used adaptive testing to estimate a student’s ability and to identify a number of student characteristics. This paper recommends content that can reinforce areas in which the student needs improvement. We applied our system at an actual education site. The group that used our recommendation module learned more effectively than the control group. By using this system, teachers will be able to monitor students closely. This enables customized learning which allows students to study effectively without necessitating the effort to search for learning materials. Customized learning will increase interest in.

A voice-controlled web browser to navigate hierarchical hidden menus of web pages in a smart-tv environment

This paper proposes a new voice web browser that can be operated in smart TV environments. Previous voice web browsers had the limitation of being run under limited conditions; for example, a list of the specific contents of a page was outputted by voice, or the user entered a search term by voice. In our method proposed in this paper, all the hierarchical menu areas on a web page are recognized and controlled with voice keywords so that page navigation according to a menu can be conveniently done in a voice supported web browser. Although many studies have been conducted on web page menu recognition, most of them provide insufficient information to recognize the hierarchical menu structure. In other words, most web pages in recent browsers showed submenus only as a result of a specific user interaction, since these previous studies had no way of recognizing or controlling the submenus. Therefore, in the web browser proposed in this study, a hierarchical menu structure, which is inserted dynamically via user interaction, is recognized and selected by voice, thus making it possible to maneuver on the web page. Furthermore, the core code of the browser is implemented in JavaScript, so it can be flexibly used not only for a web browser on Smart TVs, but also as functional extensions of existing web browsers in a PC environment.

HOTDOL: HTML Obfuscation with Text Distribution to Overlapping Layers

The purpose of this paper is to propose a method for protecting HTML content from unauthorized reuse. A document protected using this method can still be viewed with a standard text editor, but it is virtually impossible to modify, and is therefore unattractive for redistribution. Existing obfuscation methods and restriction codes that limit certain activities in the web browser are less effective due to fundamental problems of circumvention. Not only does our method address these problems using internal reconstruction of obfuscated text, but it also maintains full accessibility to web content. We begin by specifying the HOTDOL (HTML Obfuscation with Text Distribution to Overlapping Layers) standard, under which a document maintains its general layout but has its internal representation of HTML transformed into illegible text. We then suggest three alternative service architectures through which HOTDOL can be applied to existing web services. We performed a quantitative evaluation in the form of an offline survey with 12 questions, confirming the protectiveness of our method relative to other methods currently in use. To assess the computational efficiency of our method, we also measured the processing and rendering times for obfuscated documents in real environments, and found them to be commercially viable.

Dynamic Arrangement of Control in a Personalized Learning System Based on User Competency

This paper proposes the e-Space Manager to reconstruct a learning environment and describes the method used to modify components. The proposed system analyzes user competencies, and the results influence which components the users can select. That is, it provides not only for the rearrangement of components but also for the modification of components according to the user’s competencies. To verify the effectiveness of the proposed system, the e-Space Manager was applied to an SQL-based learning system. Users chose the optional learning contents that they wanted. It was confirmed that information about the learner’s competency influenced the learning components. A survey of users showed in a higher satisfaction rating for the proposed study system than for the previous system. Using this system, tutors not only could analyze the learners’ abilities precisely based on their competencies but also could recommend study areas and provide a customized environment to each individual.

Enhancing the Assessment Item Rearrangement with Bin Packing Algorithms

Online test system has been used widely in various distance learning and training systems. As more and more intelligent online systems appeared, lots of teachers could generate items and tests by diverse e-learning tools. An assessment test contains a set of items. There exist only two typical test delivery approaches supported by most online test systems. One is that a group of items are represented in a single page, and other one is that an item is presented in a single page. In case of the former approach, we need to discuss about how many items should be represented in a single page. In other words, space utilization is an important performance measure of test delivery engine. Firstly, this paper describes the design and implementation of the test generation module and the test delivery engine. And, we apply bin packing algorithms to typical items rearrangement method for responsive test sheet of multiple screen sizes. It enables each item not to be divided into two or more pages and each page to be filled efficiently with items. Lastly, our experiment shows our proposed method based on Best-Fit algorithm is superior to other algorithms with the number of pages spent.

Mining association rules of optional courses for course coordinator

In the aspect of the faculty, a course coordinator plays a significant role in managing the curriculum and counseling students on academic matters and fostering their progress in the course. However, the course coordinator cannot afford to advise students on which fields of their faculty fit them and which courses they have to take. We searched for relationships between subjects, using association rules, by mining data about the courses already taken by students, and compared these to existing course trees. And we wish to do so that can utilize to taking a course tree enactment or when revise reference data using subject information that association rule through this research. This information could be used for updating course trees.