Research On Agricultural Web Information Credibility Evaluation Method Based On Information Content

With the popularity of network technology,information technology rapid development, also agriculture in the process of social informatization gradually realize agriculture informatization.The main body of agriculture are farmers, but they are unable to distinguish the reliability of all kinds of information in network beacause of their limited scope of knowledge and without financial ability in the agricultural information service.In this paper,focusing on these problems in the process of agricultural information service, we study the academic problem about how to evaluate the credibility of agricultural network information.The major tasks of this research include four aspects as follows:(1) Aiming at the problem of without considering the words in the page's different position for traditional TF-IDF theme extract method,we put forward TF-IDF method of based on words position weight to extract theme of agriculture web information,and verified by the experiment in this paper the improved method to extract the theme of the accuracy is higher than the traditional method of TF-IDF, extraction effect is ideal;(2) Facing the difficult parts that candidate web doesn't consider the credibility during the stage of looking for candidate web pages,we propose agricultural web information credibility evaluation method based on the content,the main building has four floors credibility evaluation index:the first layer judge the authority of web page,for there is no authoritative classification and quantitative standards,this paper designed a website authority for this weight table,to distinguish between different web page authority effect is better;The second judge timeliness of web pages and put forward a new method that establishment of attenuation function within a specific time,it can better reflect the effects of timeliness of agricultural network information credibility;The third layer of judging the correlation of web pages,we generate word frequency vector of each candidate web with the introduction of VSM model,and calculate candidate web content and the relevance of the keyword;The fourth floor judge the influence of the web pages,in combination with two aspects of web links and user behavior,we can better quantify influence the size of the web pages by introducing the web site PR,Page View and Time on Page;(3) Setting different theme to reflect the number of the search term and the connection between the theme relevance,the results show that candidate web topic relevance average is 77.4% when we select four search terms,the results for the optimal;(4) Establishing search engine natural ordering,lack of correlation between sorting,and the sorting of the evaluation methods based on the content to verify this candidate web credibility.Natural ordering credibility gap value distribution;Lack of correlation between sorting put some information that nothing to do with the topic content arrange on the top position;This paper method of sorting can filter the topic content related and high credibility web page and provide users at first,which showed that the article's evaluation method based on the content is effective and practical to evaluate agricultural web information reliability.