Today, how to evaluate and improve website link structure has become a crucial issue for website design and maintenance. In this paper, we propose a novel approach for evaluating and improving website link structure based on user behavior instead of complex semantic analysis. By considering both authors’ intention and users’ preference, the evaluation and improvement strategy is relatively reasonable. By dividing the whole website into many web page pairs and calculating Association Degree and Convenience Degree of them, the complex analysis to website link structure is turned into analyzing some important partial link structures (web page pairs), which reduce the complexity of the problem greatly. We evaluate and improve website link structure in three levels: hyperlink, web page and the whole website. Experiments show satisfactory results. A prototype system for website link evaluation based on user behavior is developed, which is effective for website maintenance.
[1]
Umeshwar Dayal,et al.
From User Access Patterns to Dynamic Hypertext Linking
,
1996,
Comput. Networks.
[2]
Oren Etzioni,et al.
Towards adaptive Web sites: Conceptual framework and case study
,
1999,
Artif. Intell..
[3]
Oren Etzioni,et al.
Adaptive Web Sites: Automatically Synthesizing Web Pages
,
1998,
AAAI/IAAI.
[4]
Ramakrishnan Srikant,et al.
Mining web logs to improve website organization
,
2001,
WWW '01.
[5]
Hiroki Kato,et al.
Discovering the gap between Web site designers' expectations and users' behavior
,
2000,
Comput. Networks.
[6]
John D. Garofalakis,et al.
Web Site Optimization Using Page Popularity
,
1999,
IEEE Internet Comput..
[7]
WebUsersMyra Spiliopoulou,et al.
A Data Miner analyzing the Navigational Behaviour of
,
1999
.