Diana Purwitasari1 Yasuhisa Okazaki1 Kenzi Watanabe1
1 Saga University, Japan
1 Saga University, Japan
Abstract: The rapid growth of the Internet and the availability of abundance information
resources make Web-based learning applications become a new means to learn. We present an
approach of generating system to build a sequence of topics from existing learning resources. The
sequence list is dened as content-based navigation to represent hierarchically structured list of
topics and sub topics which is created without relying on human power. We propose use of data
mining techniques to extract topics, and then followed by use of hypergraph partitioning along
with agglomerative clustering to produce hierarchical list of topics. Another issue is an emphasis
to the term of adaptive in Web-based learning as adaptive information ltering which is to nd
items that are relevant to user interests. We interpret that content-based navigation should adapt
when context of the user interests changes like in a time user clicks a topic. Navigation generating
system makes an inquiry using feature terms within current Web page as search keywords of
selected topic when users click any topic. Our searching process applies combination ranking
factors of link analysis and content analysis which still retaining spatial information of search
keywords. Afterward, inquiry results become new resources in which the system will produce the
next suitable navigation for a collection of more focused subjects. In this paper we also show
evaluation of our selected methods for generating navigation.
resources make Web-based learning applications become a new means to learn. We present an
approach of generating system to build a sequence of topics from existing learning resources. The
sequence list is dened as content-based navigation to represent hierarchically structured list of
topics and sub topics which is created without relying on human power. We propose use of data
mining techniques to extract topics, and then followed by use of hypergraph partitioning along
with agglomerative clustering to produce hierarchical list of topics. Another issue is an emphasis
to the term of adaptive in Web-based learning as adaptive information ltering which is to nd
items that are relevant to user interests. We interpret that content-based navigation should adapt
when context of the user interests changes like in a time user clicks a topic. Navigation generating
system makes an inquiry using feature terms within current Web page as search keywords of
selected topic when users click any topic. Our searching process applies combination ranking
factors of link analysis and content analysis which still retaining spatial information of search
keywords. Afterward, inquiry results become new resources in which the system will produce the
next suitable navigation for a collection of more focused subjects. In this paper we also show
evaluation of our selected methods for generating navigation.
Download File
0 comments:
Post a Comment