Research On User Information Demand Of Recognition Based On Web Log Mining

Today, information explosion and disorientation is one of the realities faced by all information users in the era. We are eager to find the information what really need by search engine from so much information. The user often can't accurately express their information needs by submitting a query words to search engines because of user's knowledge, background, and the environment in a variety of factors. Some scholars studied the user's behavior based on the search engines, they expect to found user's interest from the user behavior. Some others got the user's information needs by questionnaire survey on the network. This paper what is different that user's behavior characteristic of information will be combined with the data mining technology to build identify the user's information demand model, in order to get the user's information needs, and expect a model for improving the efficiency of the search engine. The paper focuses on the information in through the user behavior characteristic to mining user log inquires, to establish the user's information demand classification model.First, the paper studied the main theory and technology of Web log mining and the user's information needs, described the theoretical basis of this study, and proposed to study the issue. Next, the paper gave a detailed description of pre-processing stage to web log data, and introduced the source of the data, data format, the cleaning of the log data conversion, user identification and other pre-treatment procedure. Then, classify the user's information search behavior for the potential user search behavior, summed up the search engine users are some basic behavioral characteristics and patterns of use of simple statistical methods. Fin ally, the paper Classify user's information needs based on the search engine, that is for the navigation information needs and information transaction information needs, and establish classification model of the user's information needs based on behavioral characteristics of user's information.