Multi-level classification technique for diagnosing osteoporosis and osteopenia using sequential deep learning algorithm

Sequential analysis techniques have brought new insights into a primary analysis of medical data for early detection of diseases. The presence of Sequential patterns in the datasets help to develop new drugs, define populations susceptible to certain types of illness and also to identify predictors of many medical related common diseases. At the same time, the accuracy of sequential data mining results, may depend upon techniques used, such as convolution of many attributes, feature selection method, handling of class imbalance, algorithm preference, and performance metrics. Sequential data extraction is more successful in detecting osteoporosis and osteopenia, both of which are life-threatening diseases that mostly affect women after menopause. Furthermore, its primary causes include small bone fractures, which may lead to mortality in the latter stages. As a result, the emphasis of this research work is on the development of a classifier using a deep learning technique for early prediction of osteoporosis and osteopenia diseases using health data set. In this paper, we propose, a novel sequential classifier, which is based on the deep convolution neural networks for performing classification on health care dataset relevant to osteoporosis and osteopenia, to increase classification accuracy. The proposed work distinguishes osteoporotic and osteopenia affected patients from a group of people, based on Bone Mass Density values. As per the result obtained, the proposed approach performs classification with improved accuracy and reduces the false positive rate in the detection of osteoporosis and osteopenia.