IIIT Hyderabad Publications |
|||||||||
|
Automating Web-Navigation Support Using a Cognitive ModelAuthors: sonal aggarwal,Herre Van Oostendorp,Bipin Indurkhya Conference: 4th International Conference on Web Intelligence, Mining and Semantics Date: 2014-06-02 Report no: IIIT/TR/2014/29 AbstractReducing cognitive overhead is one way to ensure fast and easy web-navigation for users. We present here an automated navigation-support tool for websites. Whenever a user visits a website for searching some information, the tool will suggest relevant links to click, where they can find desired information on that website. The tool is based on a new cognitive model CoLiDeS++Pic, which combines two related models, CoLiDeS+ and CoLiDeS+Pic. The model computes semantic similarity between the user goal and the website information using latent semantic analysis technique. It also takes into account the path adequacy, performs appropriate backtracking if required, and uses semantic information from pictures. In the current implementation, the tool is not fully automated because the semantic features of pictures have to be obtained manually. To evaluate the effectiveness of the tool, we conducted an experiment with tool support and multi-tasking as independent variables. Statistical analysis showed a significant positive impact of the tool support for time needed to perform search tasks, disorientation in navigation and task-accuracy. The navigation performance in time and disorientation of the users and answering questions are improved when provided with the tool support. Multi-tasking had no effect on time needed or answering questions, but we observed an unexpected positive effect on disorientation, suggesting that perhaps the participants were more motivated to perform primary task due to the assigned secondary task (monitoring a video at the same time), which resulted in less disorientation. Full paper: pdf Centre for Software Engineering Research Lab |
||||||||
Copyright © 2009 - IIIT Hyderabad. All Rights Reserved. |