Bonfring International Journal of Data Mining
Online ISSN: 2277-5048 | Print ISSN: 2250-107X | Frequency: 4 Issues/Year
Impact Factor: 0.245 | International Scientific Indexing(ISI) calculate based on International Citation Report(ICR)
Intelligent Health Care Data Analysis System Using Swarm based Optimization Algorithm
Abstract:
Large amounts of clinical data present both opportunities and challenges in modern healthcare. Effective data analysis improves patient care, early disease detection, and personalised treatment. This paper proposes an Intelligent Health Care Data Analysis System employing PSO and neural networks to improve disease detection. Missing values, normalisation, and feature selection ensure high-quality analysis inputs. Disease classification and prediction from clinical data are improved by PSO optimising neural network weights and biases. This method works on Hepatitis, Wisconsin Breast Cancer, and Cleveland Heart Disease datasets with and without missing values. PSO-optimized neural networks outperform regular neural networks in accuracy and robustness. Scalable and adaptable for real-time diagnostic applications, this intelligent system may improve healthcare delivery by enabling more precise and timely interventions. Using neural networks and advanced swarm-based optimisation, medical data analytics has improved patient outcomes and healthcare system efficiency. Particle Swarm Optimization-based Neural Network (IPSONN) diagnoses clinical illnesses better. IPSONN was created using K-means clustering and IPSO-trained Neural Networks. K-means clusterings optimal cluster number determines the hidden layers neurones. Cluster approximation error determines best clusters. Mean square error calculates hidden and output layer weights in QPSO. The classifier was tested with four clinical datasets from the University of California, Irvine (UCI) machine learning repository: Pima Indian Diabetes, Hepatitis, Wisconsin Breast Cancer, and Cleveland Heart Disease.
Keywords: Health Care, Heart Disease, Machine Learning, Disease Detection.
Volume: 14 | Issue: 2
Pages: 1-7
Issue Date: August , 2024
DOI: 10.9756/BIJDM/V14I2/BIJ24011
 |
|