Smart System for Lung Disease Early Detection

Tuberculosis and Pneumonia are two of the ten deadliest diseases in Indonesia, and Chronic Obstructive Pulmonary Disease (COPD) ranks fourth as the leading cause of death in the world. In fact, Indonesian people still lack awareness of COPD. However, some conditions of body organs such as lungs can...

Full description

Saved in:
Bibliographic Details
Published in:Journal of physics. Conference series 2018-12, Vol.1140 (1), p.12035
Main Authors: Hasanah, N, Arifin, F, Irmawati, D, Muslikhin
Format: Article
Language:English
Subjects:
Acoustics
Bacterial diseases
Chronic obstructive pulmonary disease
Decision making
Image classification
Image processing
Optimization
Organs
Tuberculosis
Waterfalls
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Tuberculosis and Pneumonia are two of the ten deadliest diseases in Indonesia, and Chronic Obstructive Pulmonary Disease (COPD) ranks fourth as the leading cause of death in the world. In fact, Indonesian people still lack awareness of COPD. However, some conditions of body organs such as lungs can be diagnosed through sounds. Although pulmonary auscultation has been widely used, technological advancement in the field of digital lung sound analysis and classification provide new possibilities that need to be explored. ANFIS is a popular branch of AI science and is widely applied for classification, prediction, and image processing. ANFIS is essentially able to implement human expertise used in a decision-making process. The objectives of this study are (1) to develop a smart system for the early detection of lung diseases using ANFIS, and (2) to determine smart system accuracy. The smart system was developed using a Waterfall model in sequence starting from analysis, design, coding, and testing. ANFIS that had been built consisted of four inputs and one output, with 81 rules. At the ANFIS development stage, several ANFIS architectures with MF type and different optimization methods had been tested. The smallest RSME value was 4.5357e-06, which was obtained from the training data testing using MF trapezoid (trapmf), hybrid optimization method, error tolerance of 0.0001, and 30 epochs.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/1140/1/012035