Indian Journal of Science and Technology

Abstract

Objectives: The present research work generally focuses on predicting diseases from the lung disease test by using data mining techniques for spirometry data. Methods/Statistical Analysis: Spirometry is used to create baseline lung function, check out dyspnea, disclose pulmonary disease, watching effects of therapies used to treat respiratory disease, calculate respiratory impairment, evaluate operative risk, and performs surveillance for occupational-relevant lung diseases. Pulmonary function tests are used to find out lung capacity, based on which the many of the lung diseases can be identified. In this research work, a combination of k-means clustering algorithm and Decision tree algorithm was developed. From the results investigation, it is known that the proposed aggregated k-means algorithm and decision tree algorithm for spirometry data is better which compared to other algorithms such as Genetic algorithm, classifier training algorithm, and neural network based classification algorithms. Findings: Existing algorithms are unable to handle noisy data and also with Failure occurrence for a nonlinear data set. It should not classify the data set based on their input attributes. Prediction is not possible for existing system. Applications/Improvement: Spirometry data which is used to predict the lung capacity using Aggregated K-means and Decision tree algorithm. Our proposed approach is evaluated for each dataset accordingly

Keywords: Decision Tree, Pulmonary Function Test Means, Spirometry Data