Classification Algorithm Of Unbalanced Datasets

With the rapid development of computer technology and the internet resource sharing, data mining becomes a hot spot in recent years. Data mining is the meaning that access to effective, innovative, hidden valuable information from massive data. Unbalanced data set classification algorithm is an important branch of data mining. In the study of unbalanced data sets classification algorithm, the main work is completed as follows:Starting from the sampling method unbalanced datasets, this article designs the constraint sample method. Firstly, this method treatments with c-means clustering method on unbalanced datasets, and then executes subsamping deal on the processed data to reduce imbalance ratio of many types of samples in the dataset. Finally, combines the constraint sample method with SVM, and classifies the unbalanced datasets.Then this article study the construction method unbalanced datasets fuzzy classification system based on genetic algorithm. Firstly, identify the initial classification system based on fuzzy clustering algorithm GK. Then this article optimizes initial fuzzy classification system using genetic algorithms to improve the accuracy of fuzzy classification system.Finally, explanatory problem of the fuzzy classification system obtained by the above algorithm is poor. So the article studies the coevolutionary algorithm of unbalanced datasets fuzzy classification system, which can optimize structure and parameters of the obtained initial fuzzy system. Improve the explanatory classification system maximizingly under the premise of ensuring accurate.