Personalized Recommendation Service Study Of Agricultural Information Based On Context-aware Integration

Agricultural information is an important part to promote China's modernization in the development of agriculture.The number of relevant information resources that generated in the process of its evolving is rapidly growing at an exponential level.However,agricultural workers with poor modern information technology or knowledge usually encounter problems like "information overloaded" and "information disoriented" while obtaining the required information resources,which is not conducive to the further work in the construction of agricultural information.With the continuous growing of China's economy,increasingly improvement of the living standards of farmers,further improvements of rural information infrastructure,it turns to be a row of effective solution to build a personalized agricultural information service system and provide users with targeted resources timely and effectively with modern information technology.This paper discusses the analysis of problems in the actual need for real and agricultural information service system.And this paper puts forward a set of research program based on context-aware integration personalized recommendation agricultural information service system to solve problems encountered in the following three core modules:building and updating of users' interest model,modeling and classification of information resource model,and personalized recommendation service.As a result,it improves the performance and accuracy of personalized information recommendation service in the agricultural information resource referral service process.The main work is as follows:This paper reviews the field of agricultural information resources integration scenarios based on personalized recommendation algorithm theoretical knowledge and analysis of agricultural information service system.From the points of participators in information services,information services contents,and information service channels,this paper found problems and needs in personalized agricultural information service and built the framework of personalized recommendation agricultural information service based integration for situational awareness.The author uses improved agricultural information resources TFIDF algorithm for feature extraction to establish a multidimensional model of agricultural information resources and agricultural information resource feature dictionary.Through the characteristics of the different dimensions of resources from the type,time and space and objects in three dimensions by classification vector space model,this paper completes automatic acquisition of agricultural information resources,modeling and classification.According to the rules established by the knowledge of the agricultural sector,this paper put the perceived user profile information into the users' interest model.With combination of explicit and implicit user acquisition and nearest neighbor collaborative filtering information of interest,this paper initialized and updated the users' interest in the two-dimensional and object dimensions.The author implements the recommended information resource model,the user interest model as well as recommendation based on contents by java and MySQL,providing solutions for the agricultural-based personalized information recommendation service integration framework for situational awareness.This paper contrasts personalized information services and traditional agricultural information services with experimental data and reflects users' interest in migration sensitivity,accuracy and precision rate recommendation information,recall and F-measure value in a great degree.Innovation of this paper focuses on the following two parts:First,with clear felling of the user's perception and situational information into user interest model which combines features of the agricultural sector knowledge,this paper proposed a scenario information for new processing methods to take advantage of pervasive computing,which allows computer users to obtain more accurate individual needs and improve the accuracy of the information recommendation service system.Second,adopt the vector classification method for agricultural information resources and user interest model,which shows both reflect cyclical,regional,dynamic,practical,scientific characteristics of agricultural information resources.Meanwhile,it can be further carried out for the feature an item refined distinction and excludes the mutual interference of the different dimensions of the same characteristics as the word so that it improves the accuracy of model representation.