Research On Algorithm For Relational Data Classification Based On Background Knowledge

Data mining could combine data analysis and complex algorithm for processing huge data sets. It offered wonderful opportunities for exploring and analyzing new data and old data. The purpose of data mining is to find unknown useful patterns in huge data sets. As a branch of data mining, relational data mining is becoming hotspots of this area. The paper has done some research on relational classification.Firstly, a criterion for classifying relationships between tables in relational databases is proposed by analyzing large number of instances. It can also give these relationships evaluations formally and semantically.Secondly, a classification algorithm for relational data based on background knowledge is proposed. It could classify relational data by constructing a relational decision tree. The algorithm takes the information gain to evaluate attributes, and it could take useful background knowledge into target tables to guide the classification by introducing the class label propagation techniques. In addition, it could support relational databases directly.Thirdly, a relational classification algorithm with users'guidance is proposed, which improves naive Bayesian classifier. It introduces users'guidance and sharply reduces searching time for a useful attribute to make users more satisfied. The classification accuracy is higher than other algorithms.Experimental results show that the algorithms proposed in this paper are more efficient than the current ones, and the anticipated results are realized.