Abstract:
Web services are getting more and more diverse and people may get confused with thou-
sands of web services returned from search engines or crawlers whenever they want to
nd a web service for their purposes. This problem has impulsed researches on propos-
ing a web service recommender system to lead users to expected items while hiding
unnecessary ones. Most of proposed solutions analyzed query strings and web service
descriptions to give recommendations. However, text based analysis mostly depends
on user's perspective, languages and notations which easily decrease the performance
of algorithms. Moreover, new published web services often have lower priorities to
be selected for recommendations. To overcome those problems, our research focuses
on constructing a recommender system based on user's interactions. We propose two
algorithms which are used for generating recommendations based on the principle of
collaborative ltering. Additionally, we propose another algorithm to assign priorities
for new comming web services to enhance web service discovery.
