Research Of User Interest Path In Web Log Based On Ant Colony Algorithm

With the rapid development of information technology, the Internet information is growing exponentially, various websites are springing generally active in our lives.For network service providers, how to dig out the massive server log file user interest model, has become a hot topic for many companies. When the ant colony algorithm for web users to access the data mining community, the web users as artificial ant ants preferred path algorithm models showcased is the path the user’s interest, according to the web user’s interest path that can adapt to different customers demand, optimize the site topology, improve system efficiency and adjust marketing strategies accordingly,improve the scientific decision-making, thereby creating greater business value for the enterprise. Practice and analysis show that the ant colony optimization algorithm for mining web users interested path MF compared to other traditional algorithm has a higher accuracy in the field of web log data mining has better application value.This study path web log user interest based on ant colony algorithm to start, the main work is as follows:(1) the use of web users interested in foraging behavior of ant colony path similarity, proposed the "intent pheromone" a new concept for the user to access web site reaction degree of interest and intention to use pheromones given based on ant swarm optimization path web log mining user interest. Namely through the pheromone global updating, calculate transition probabilities, transition probability threshold settings and other steps to finally get user interest path. Experiments show that the method is feasible and can accurately reflect the user’s interest path.(2) by pretreatment of the client log data, obtaining mouse scrolling, relatively user browsing time, page hits to reflect the relative frequency of interest-critical information from the log data. Users browsing time which means the relative proportion of the total page browsing time accounts for the average access time for all pages, page hits, said the number of page views cent of all page visits proportion mouse scroll mouse scroll represents the relative number of web pages when users browse representing the number of times all the pages of the mouse scroll proportion experiments show that the pheromone consisting of three to more accurately represent the interest of the users of the site.(3) the ant colony algorithm?and?determines the pheromone concentrationand the degree of influence choice and preference, the paper redefines them, through,the adaptive heuristic function to adjust to the new definition can not only accelerate the convergence speed, and local convergence and avoid premature, experiments show that this method improves the search efficiency of the algorithm.