Web Information Credibility Research Based On Information Fusion

:When users find the information they need in the network, the most used tool is the search engine. But the search engine result does not allow users with particular satisfaction as a business tool. Of course, the user can select one by one to find out useful results they need. But the direct use of search engine results will increase the workload for using in information fusion, and may lead to inaccurate results. So the thesis proposes to evaluate the web information credibility based on information fusion.Through analysing the lack of search engines on the information credibility, then according to the characteristics of web information and requirement of information fusion, it is concluded that the most important effect of Web information credibility is webpage correlation. The thesis does research on information credibility, establishes the information credibility evaluation, puts forward the calculation method of credibility.The main contents of the thesis are as follows:1. Through Analysing the more using of computation webpage related degree at present, according to the characteristics of information fusion, the thesis uses Frequency position weighted sorting algorithm. According to the shortcomings of the original algorithm, an improved vision is proposed. We introduce TextRank model to extract keywords, and add the weight of word’s relocation itself to the original TextRank model, it introduces the concept of semantic, and also consider the location of the words on the page, that improves the accuracy of keyword extraction. When Calculate the correlation, the words in the thematic words extraction steps to calculate words weight is introduced in the formula, and also considers the semantic similarity between keywords and query words. When calculating correlation, we consider the semantic relations between words and words, frequency and position of words, that makes the calculation more accurate.2.Summarize the information credibility of the existing evaluation, through the analysis of search engine evaluation in the lack of credibility, we structure the evaluation index system of information credibility, from the three aspects of authority, importance, correlations to evaluate reliability, selected the most relevant indicators of each aspect, put forward to calculate the reliability formula. The formula to calculate the reliability is considered some of the most influential and the most objective index, the credibility of the value is closest to the information fusion.3.We design and implement a system of the credibility evaluation to verify the effectiveness of the method proposed in the thesis, and the results are analyzed. The results show that the proposed algorithms are effective and practical.