Search

Support vector machines (SVMs) are known to be useful for separating data into two classes. However, for the multiclass case where pairwise SVMs are incorporated, unclassifiable regions can exist. To solve this problem, Fuzzy support vector machines (FSVMs) was proposed, where membership values are assigned according to the distance between patterns and the hyperplanes obtained by the "crisp" SVM. However, they still may not give proper decision boundaries for arbitrary distributed data sets. In this paper, a density based fuzzy support vector machine (DFSVM) is proposed, which incorporates the data distribution in addition to using the memberships in FSVM. As a result, our proposed algorithm may give more appropriate decision boundaries than FSVM. To validate our proposed algorithm, we show experimental results for several data sets. [ABSTRACT FROM AUTHOR]

Companies must have tools to manage effectively their huge engineering data. So we developed a data mining system based on GP which can be one of the components for the realization of the utilization of engineering data. The System can derive and extract necessary empirical formulas to predict design parameters in ship design. But we don't have enough data to carry out the learning process of genetic programming. When the learning data is not enough, not good result such like overfitting can be obtained. Therefore we have to reduce the number of input parameters or increase the number of learning and training data. In this paper we developed the improved data mining system by Genetic Programming combined with Self Organizing Map (SOM) to solve these problems. By using this system, we can find and reduce the input parameters which do not influence on output less, as a result of this study we can solve these problems. [ABSTRACT FROM AUTHOR]