Data Classification analysis using Distance Function in Learning Vector Quantization Algorithm

Dushyant Purohit
Compucom Institute of technology & Management, Jaipur , IN

Yogendra Sharma
Compucom Institute of technology & Management, Jaipur , IN

Jitender Kumar
Compucom Institute of technology & Management, Jaipur , IN

Ritu Vashistha
Compucom Institute of technology & Management, Jaipur , IN

Keywords: Clustering, Artificial Intelligence, Neural Networks, Self Organizing feature Map ,Learning Vector Quantization, Voronoi vectors, Area under the curve(AUC)

Abstract

Artificial Neural Networks (ANNs) are human invented information processing objects, highly parallelized dynamic system that yields output by reapplying output as input. They have confirmed to be beneficial in solving problems which cannot be solved by using heavy algorithmic procedures in area like mining,categorization of large data, image processing etc. Proposed work focus on prediction of diseases like Cancer, Diabetes, patient hospitalization etc. Proposed work is to train the data with a new approach having combined supervised and unsupervised mode and get results with new conceptualize approach on test data. This approach is experimented by Learning Vector Quantization (LVQ) algorithm because of its characteristic of multimode working and uses heuristic approach for searching. Multimode working includes first is unsupervised step and another is supervised. Diabiteas dataset is retrived from UCI learning respiratory to test the proposed idealogy on algorithm . The program is implemented on Matlab and result is shown using AUC (Area under the curve). Our approach is to improving accuracy in such a way that the performance may high and computational cost remain less.

How to Cite

References

NA