An intelligent system for clustering using hybridization of distance function in learning vector quantization algorithm

Artificial Neural Networks (ANNs) are human made information processing artifacts, and grown up vast in two-three decade. Neural Networks are highly parallelized dynamic system which accept output response as input and produce output. They have confirmed to be extensively beneficial in solving those problems which cannot be solved by using algorithmic procedures which are considered to be conventional, or possess a highly complex solution in area like prediction and classification of large data, image processing etc. Automation using ANNs in medical sector used to are carry out diagnoses of disease and prediction of disease etc. Proposed work focus on hybridization of distance functions for computing distance and increasing the prediction of diseases like Cancer, Diabetes, cancer. Proposed work is to train the data with hybrid function and get results with hybrid function on test data. We have used Supervised Learning Vector Quantization (LVQ) to apply this hybrid function because of its characteristic to work in two steps and used heuristic search technique. First is unsupervised step and another is supervised. We have also shown comparison tables in proposed work which distance function works best on the dataset by showing area under the curve measuring on MATLAB. The data sets for providing learning and examining are taken from medical science and increasing effectual identification of diseases like Cancer, Diabetes by using data classification techniques, thereby modifying ANN algorithm to enhance their performance. We have also test the proposed algorithm on randomly generated dataset to maintain sustainability. of The whole experimental work is simulated using MATLAB 7.7.0 and showed that increasing the accuracy of data by using proposed hybridization of distance function.