Using Blood Indexes to Predict Overweight Statuses: An Extreme Learning Machine-Based Approach.

Area of Science:

• Biomedical Engineering
• Machine Learning in Healthcare
• Public Health

Background:

• Rising global overweight rates pose significant health risks, including hypertension, diabetes, and heart disease.
• Accurate identification of overweight status is crucial for preventative healthcare strategies.

Purpose of the Study:

• To develop and validate a novel machine learning-based method for detecting overweight status using blood and biochemical measurements.
• To identify key biochemical indicators associated with being overweight.

Main Methods:

• An extreme learning machine (ELM) model was developed and trained on data from 225 overweight and 251 healthy subjects.
• The model's performance was evaluated using accuracy, sensitivity, specificity, and Area Under the Receiver Operating Characteristic Curve (AUC).
• Feature selection analysis was performed to identify significant biochemical markers.

Main Results:

• Significant differences in blood and biochemical indexes were observed between healthy and overweight individuals (p < 0.01).
• Key biochemical markers identified include creatinine, hemoglobin, hematokrit, uric acid, red blood cells, high-density lipoprotein, alanine transaminase, triglyceride, and γ-glutamyl transpeptidase.
• The ELM model demonstrated high accuracy in detecting overweight status.

Conclusions:

• The proposed machine learning approach using blood and biochemical data offers a promising and accurate method for identifying overweight individuals.
• The identified biochemical markers provide insights into physiological changes associated with overweight status.
• This technique can support early detection and intervention for overweight-related health risks.