Research and Implementation of Clustering Analysis Algorithms Based on I-MINER

I-MINER is convenient to establish data mining model and embed other data mining models with I-Miner. DBSCAN algorithm can achieve clustering of any shape of dataset, Fuzzy C-Means is suitable for the dataset which is uniformly distributed around cluster centers and CABOSFV algorithm can be a good clustering for high-dimensional dataset (such as WEB data). In this thesis, DBSCAN, Fuzzy C-Means and CABOSFV clustering analysis algorithms are embedded into I-Miner to enormously satisfy users´ needs, establish data mining model and support production decision-making, besides, the three mining models are compared. Through three mining models, mining and comparative analysis are made for examples to get the advantages and disadvantages of the three clustering algorithms.