A Trend-based Prediction System for Web User Behavior

Since web applications make great progress, the latency of Internet owing to the network bandwidth becomes an urge problem in the cyber world. It is very important to deliberate on how to construct a prediction model to predict web users traveling path for adapting the website structure and improving the website performance. A trend based prediction model without extra information is proposed in this paper to generate prediction models with a sequence of pages for a proxy server prefetcting the suitable pages. The trend similarity is the core of our proposed model which considers not only the page similarity but also position similarity. Two measures include page correctness rate and order correctness rate are proposed to evaluate accuracy of our prediction system. Key-Words: Trend similarity, Prediction system, Web mining, User behavior, Sequence mining