SIMILARITY-BASED FEATURE SET SUPPLEMENTATION FOR CLASSIFICATION

摘要

One group of associated first item of neighbours' project has been identified the first item for being based in part on fisrt feature collection, the associated feature set of project of neighbours' group of each project neighbours. The feature set of supplement is generated first entry and is combined the first feature set and characteristic set and this group of adjacent article based on one group of neighbours' item response being identified. One group of associated first entry of classification score is based on complementary features collection, each classification score value, and various classification scores indicate that first entry may belong to a kind of article.