USER NAVIGATION DESIGN WITH DOCUMENT CLUSTERING FOR DIGITAL LEARNING RESOURCE IN INDONESIAN (CASE STUDY: Q-JOURNAL PT TELKOM INDONESIA)

AGUSTINA WULANDARI

Informasi Umum

Kode

16.05.015

Klasifikasi

006.312 - Data mining

Jenis

Karya Ilmiah - Thesis (S2) - Reference

Subjek

Data Mining

Dilihat

225 kali

Informasi Lainnya

Abstraksi

Information technologies are growing rapidly today. Those are also used by content providers to deliver a web based application that can help educational institutions to maintain their knowledge resources such as e-journal. There are many e-journal applications have been built that we can find easily, but most of them are English based journals. Among the few e-journal applications in Indonesian, it was rarely well maintain. Q-Journal, developed by PT Telkom Indonesia, is an integrated digital platform for discovery and publishing research result such as journals. To be a competitive application, e-journal system must provide a better experience to users by creating a learning journal contents system. Learning contents mean that the system has capabilities to understand what users need by learning their behavior and store it as form of user profile. Once user profile created, the system will suggest the other contents which are match with their profile for the next visit or search.

Those capabilities could only be implemented by using data mining techniques. Combine with suitable algorithms and approaches, the application could deliver an effective and efficient both for the system itself and the user. In this research, Hierarchical Agglomerative Clustering (HAC) method capable to provide a good accuracy in performing topic clustering. So it has generated learning content as a good user navigation system for digital learning resources. The validation process is calculated by cophenetic correlation coefficient value (CPCC).

As a result, average value performance measurement of hierarchical clustering result produced by CPCC is 0.91 that indicates the clustering result have a good performance. The implementation of the propose design for user navigation has proved to be able providing a better result for e-Journal existing system. The structure of hierarchy topic-subtopic can be organized with keywords extraction approach so it become a good user navigation. The propose system can optimize facilitating contextual browse by considering user information which is adopted from users' interest. Implementation of user navigation design makes easier for user getting information proper to their interests. User navigation generating system is considered to allow a user to find a learning path for adjustment the user's interest on a particular learning topic.