期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2016
卷号：7
期号：2
页码：820-823
出版社：TechScience Publications
摘要：Bone disease prediction is the most required task inthe real world environment which needs to process well toattain the improved prediction of bone diseases. For analysingthe bone disease related data sets, various data miningapproaches needs to process to obtain the improved predictionof the disease. One of the main data mining approaches whichis used for predicting the risk factor of bone diseases aremachine learning algorithm. Machine learning is a subfield ofcomputer science that evolved from the study of patternrecognition and computational learning theory in artificialintelligence. Machine learning explores the study andconstruction of algorithms that can learn from and makepredictions on data. The main disadvantage of existingmethod is that, it will not produce accurate result of bone lossrate due to limited number of features. This drawback issolved in our proposed method by using the relevance basedfeature selection which tries to predict the result with lessavailability of feature. Risk factors are fine-tuned using Deepbelief network. In this algorithm 2 phases of process arepassed out. It includes Pre-training and fine tuning. In thepre-training phase, most important risk factors with modelparameters are used to calculate contrastive divergence and itminimizes the record size. In the fine tuning phase comparisonis made with the results achieved in the previous phase withthe ground truth value g1 and again the same comparisondone with ground truth value g2, were g1 is refer to asosteoporosis and g2 is refer to as a bone loss rate. The finalresults are applied to confusion matrix to describe theperformance of classification model based on the comparisonresults to calculate Accuracy. The experimental result of thiswork proves that the proposed methodology of this workprovides performance improvement than the existingmethodology in terms of more accuracy.
