LUNG DISEASES IDENTIFICATION WITH HYBRID GENETIC- MLP APPROACH
RAMANDEEP KAUR *
Department of Computer Science and Engineering, CT Institute of Engg. Mgt. & Tech, Jalandhar, India
PRINCE VERMA
Department of Computer Science and Engineering, CT Institute of Engg. Mgt. & Tech, Jalandhar, India
*Author to whom correspondence should be addressed.
Abstract
Data Mining is the process of extracting or mining information from the large volumes of data. Classification is a data mining task, used to classify data among different predefined clusters. The paper proposes a new classifier utilizing MLP approach by maintaining the clusters based on similar feature with distance formula. With using cluster based method MLP approach can handle large data and increases the accuracy. This technique has been applied for automating medical diagnosis. This paper analyzes the lung images (i.e. CT-scan images) for identifying and classifying them among the various lung diseases (i.e. bronchitis, emphysema, pleural effusion or normal) by emphasizing on the problematic area.
Keywords: Data mining, classification, multilayer perceptron