Feature contribution means that what features actually participates more in g rouping data patterns that maximizes the system's ability to classify object instances. In this paper, modified K-means fast learning artificial neural network (K-FLANN) was used to cluster mult idimensional data. The operation of neural network depends on two parameters namely tolerance (δ) and vigilance (ρ). By setting the vigilance parameter, it is possible to extract significant attributes fro m an array of input attributes and thus determine the principal features that contribute to the particular output. Exhaustive search and Heuristic search techniques are applied to determine the features that contribute to cluster data. Experiments are conducted to predict the network's ability to extract important factors in the presented test data and comparisons are made between two search methods.
K-means fast learning artificial neural network (K-FLANN) algorithm begins with the initialization of two parameters vigilance and tolerance which are the key to get optimal clustering outcome. The optimization task is to change these parameters so a desired mapping between inputs and outputs (clusters) of the K-FLANN is achieved. This study presents finding the behavioral parameters of K-FLANN that yield good clustering performance using an optimization method known as Differential Evolution. DE algorithm is a simple efficient meta-heuristic for global optimization over continuous spaces. The K-FLANN algorithm is modified to select winning neuron (centroid) for a data member in order to improve the matching rate from input to output. The experiments were performed to evaluate the proposed work using machine learning artificial data sets for classification problems and synthetic data sets. The simulation results have revealed that optimization of K-FLANN has given quite promising results in terms of convergence rate and accuracy when compared with other algorithms. Also the comparisons are made between K-FLANN and modified K-FLANN.
In this paper, fuzzy c-means algorithm uses neural network algorithm is presented. In pattern recognition, fuzzy clustering algorithms have demonstrated advantage over crisp clustering algorithms to group the high dimensional data into clusters. The proposed work involves two steps. First, a recently developed and Enhanced Kmeans Fast Leaning Artificial Neural Network (KFLANN) frame work is used to determine cluster centers. Secondly, Fuzzy C-means uses these cluster centers to generate fuzzy membership functions. Enhanced K-means Fast Learning Artificial Neural Network (KFLANN) is an algorithm which produces consistent classification of the vectors in to the same clusters regardless of the data presentation sequence. Experiments are conducted on two artificial data sets Iris and New Thyroid. The result shows that Enhanced KFLANN is faster to generate consistent cluster centers and utilizes these for elicitation of efficient fuzzy memberships.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.