Adaptive Link Sequence Recommendation based on Ant Colony Metaphor

Abstract – Website visitors often prefer different link sequences depending on their goals and changing interests to reach at desired pages in a website. There can be more than one alternate path to reach at any web page. When browsing large websites, users often depart from main subject temporarily and pursue digressive paths within websites, prolonging navigation time. In order to serve the website visitors better, this paper proposes Adaptive Link Sequence Recommendation (ALSR) approach which identifies target pages based on group of user access patterns, displays most prominent K target pages and recommends adaptive link sequence to them. Adaptive link sequence refers to selecting the order of links by aggregating surfing activities of individual users into a collective preference and their temporal perceptions into a collective perception. Adaptive link sequence computation is inspired from the foraging behavior of biological ants. Also, pheromone updating strategy known from ant algorithms for computing the weights on the edges of the co-occurrence graph containing web pages as nodes is applied to assess user access patterns. A prototype website is designed and the proposed approach is tested. Experiments show that the approach is efficient and allows users to arrive at the most prominent target pages quickly. Keywords- Ant colony metaphor, pheromone updating, co-occurrence graph, website navigation I