The Research Of Information Evaluation Based On Sources Dependence

Every data sources will provide many data, which used to reflect the information of aparticular fact, such as an online bookstore, it will provide a lot of books’ information,including the author, title, the press and the price of the book. Then these values iscorresponding to a data set, however, there is no direct relations between the facts and thetruth."Facts" does exist, and may not exist. And "truth" must exist in the facts, facts ismisleading sometime. Different data sources will often provide some conflicting values,which is reliable, in order to offer users the high quality of the data value, the key is the datamanagement applications can solve data conflicting problem, and to extract the right values.We usually think that the true values will be provided by most data source, and there islittle source which give the wrong data value. According to this view, we use votesalgorithm,which was that the most data sources provide the value as the correct value, but avalue is whether correctly or not, that will be copied by a lot of other data. In such a datainformation times, extracting accurate data will be tricky. So the truth finding should be puton more extensive attention. In this paper,we realize data classification, the specificimplementation strategy of conflicting data processing, we also describe the detailed processof capture the characteristics of a value data which is constantly changing, and observe thesimilarity between the value relations and a series of logical structure on data, accuratelylocate reliable data value of particular object.