Identifiable Temporal Feature Selection via Horizontal Visibility Graph Towards Smart Medical Applications

Area of Science:

• Medical Informatics
• Data Science
• Biomedical Engineering

Background:

• The Internet of Medical Things (IoMT) generates vast amounts of time series sensor data.
• Effective analysis of this data is crucial for detecting abnormalities and advancing smart healthcare.

Purpose of the Study:

• To develop a novel temporal classification model for disease diagnosis using time series data.
• To introduce an identifiable temporal feature selection method based on horizontal visibility graphs.

Main Methods:

• Utilized horizontal visibility graphs (HVG) for time series representation.
• Developed a temporal classification model for disease diagnosis.
• Implemented feature selection techniques for improved accuracy.

Main Results:

• Demonstrated superior accuracy and efficiency compared to existing methods on benchmark datasets.
• Validated the effectiveness of the HVG-based approach for time series classification in disease diagnosis.

Conclusions:

• The proposed HVG-based temporal classification model offers a promising solution for disease diagnosis in smart healthcare.
• The method provides accurate and efficient analysis of IoMT time series data, with released code for community use.