A Study In Applications Of Neural Network In Data Mining

The data in Data Mining is often noisy, non- linear and unorderly. Neural Network has the advantage to process these data. At the same time, Neural Network's processing need a great deal of data which is used for producing sufficient learning and test set in order to learn and evaluate effectively the performance of Neural Network. The data mentioned above could be provided by the data mining, which is based on the data warehouse or big database. Based on their own advantages and the cooperation with many other technologies, a Study in Applications of Neural Network in Data Mining is taken out.Data classification and prediction are important mining technologies and have been used widely. Nowadays, many classification methods and some prediction technologies have been put forward. K-Nearest Neighbor (K-NN) Classifiers is a typical one. Facing the massive volume and high dimensional data, how to build effective and scalable algorithm for data mining is one of research directions of data mining. This paper implemented an improved K- Nearest Neighbor algorithm in which Fuzzy Adaptive Resonance Theory (Fuzzy ART) and BP Neural Network are applied in K-NN classification to make a new algorithm. Fuzzy ART clustering is carried out to select the subset of the training set and BP neural network calculated the attribute's weight, which can reduce the volume of the training set and lead to computational efficiency.In the process of data mining on the Multi-dimensional discrete time series, the time series needs to be transformed into symbolic sequences. This paper proposes an efficient method for the symbolization of time series. This method applies Fuzzy ART to select the subset of the training set, which means clustering the Multi-dimensional discrete time series and symbolizing them. Before clustering, dependence analysis by SQL Sever is used to reducing the no or less attributes, which helps to keep the accuracy of clustering.. At the same time, the effectiveness of these methods is proved by an experiment about Huang Heshan tunnel's traffic flow.Matlab experiments of the two algorithms showed they had efficiency.