Multiple Classifier System Based Lung Disease Prediction with Spark Framework in Big Data

Main Article Content

V.Durgadevi, et. al.

Abstract

Technologies related to Big Data are potentially effective for transforming healthcare information and have developedseveral industries. Moreover, as cost is reduced, numerous lives are saved and the results are improved. Lung disease causes more death worldwide. The death rate can be reduced when detection is done at early stages but signs and symptoms are not clear in Lung disease at that stage. Hence, preventing or predicting is relatively difficult. This paper focuses on developing a prediction model for diagnosing lung disease which employs multi-structure integrated dataset.With the big data framework for healthcare and accurately predicting lung problems at the earlier stages, using machine learning approaches are considered to be the best. In this research work, a sequence of machine learning methods along with apache spark architectureis proposed for effective data classification and predicting therisk level of disease appropriately. The proposed algorithm is named as Spark framework with Multiple Machine Learning Classifier Algorithm (SMMLCA) which uses Naïve Bayes and J48 Classifier. The proposed approach is compared with two standard methods namely Convolutional Neural Network based Multimodal Disease Risk Prediction (CNN-MDRP) algorithm and Recurrent Machine Learning (RML)-based prediction modelsinterms of accuracy, precision, recall, F1-meassure, ROC and AUC. It is found that the proposed SMMLCA achieves 85.4% of accuracy, 84.2%of precision, 74.2% of recall, 71.4% of F1-measure, 75%of AUC and 61.4% of ROC

Article Details

Section
Articles