Pattern Recognition Based on Fuzzy Set and Genetic Algorithm

In this paper, we consider a soft computing approach to pattern classification. Our basic tools for soft computing are fuzzy relational calculus (FRC) and genetic algorithm (GA). We introduce a new interpretation of multidimensional fuzzy implication (MFI) to represent our knowledge about the training data set. We also consider the notion of a fuzzy pattern vector to handle the fuzzy information granules of the quantized pattern space and to represent a population of training patterns in the quantized pattern space. The construction of the pattern classifier is essentially based on the estimate of a fuzzy relation Ri between the antecedent clause and consequent clause of each one-dimensional fuzzy implication. For the estimation of Ri we use floating point representation of GA. Thus a set of fuzzy relations is formed from the new interpretation of MFI. This set of fuzzy relations is termed as the core of the pattern classifier. Once the classifier is constructed the non-fuzzy features of a test pattern can be classified. The performance of the proposed scheme is tested on synthetic data. Subsequently, we use the proposed scheme for the vowel classification problem of an Indian language. Finally, a benchmark of performance is established by considering multiplayer perception (MLP), support vector machine (SVM) and the present method. The Abalone, Hosse Colic and Pima Indians data sets, obtained from UCL database repository are used for the said benchmark study. This new tool for pattern classification is very effective for classification of patterns under vegue and imprecise environment. It can provide multiple classification under overlapped classes.

Frequent Pattern Mining Based on Pattern Space Division in Map/Reduce Cluster

By means of pattern space division and based on Map/Reduce, the problem of processing the many-to-many corresponding relationship between the data set and the patterns set is converted to the problem of processing the many-to-many corresponding relationship between the data subsets and the pattern subspaces associated with the frequent 1-itemsets. Thus, the scale of the intermediate key/value pairs set is reduced so dramatically that the problem of single Map node bottleneck which results from combinatorial explosion of candidate patterns space is avoided. Over three rounds of Map/Reduce tasks, the pattern space is constructed and divided, the filtering rules is established and employed, father more, the mining of frequent patterns is realized in each pattern subspace independently. By making the best of both the universal trait of the entire pattern space and the individuality of each pattern subspace, the optimized non-recursive algorithm is designed and implemented to improve the efficiency of mining phase.

Supervised and Unsupervised Information Granulation

In this study, we are concerned with information granulation realized both in supervised and unsupervised mode. Our focus is on the exploitation of the technology of hyperboxes and fuzzy sets as a fundamental conceptual vehicle of information granulation. In case of supervised learning (classification), each class is described by one or more fuzzy hyperboxes defined by their corresponding minimumand maximum vertices and the corresponding hyperbox membership function. Two types of hyperboxes are formed, namely inclusion hyperboxes that contain input patterns belonging to the same class, and exclusion hyperboxes that contain patterns belonging to two or more classes, thus representing contentious areas of the pattern space. With these two types of hyperboxes each class fuzzy set is represented as a union of inclusion hyperboxes of the same class minus a union of exclusion hyperboxes. The subtraction of sets provides for efficient representation of complex topologies of pattern classes without resorting to a large number of small hyperboxes to describe each class. The proposed fuzzy hyperbox classification is compared to the original Min-Max Neural Network and the General Fuzzy Min-Max Neural Network and the origins of the improved performance of the proposed classification are identified. When it comes to the unsupervised mode of learning, we revisit a well-known method of Fuzzy C-Means (FCM) by incorporating Tchebyschev distance using which we naturally form hyperbox-like prototypes. The design of hyperbox information granules is presented and the constructs formed in this manner are evaluated with respect to their abilities to capture the structure of data.