Related Experiment Video
Updated: Jun 24, 2025
12:02
Automated Modular High Throughput Exopolysaccharide Screening Platform Coupled with Highly Sensitive Carbohydrate Fingerprint Analysis
Published on: April 11, 2016
11.5K
Utilizing machine learning algorithms for precise discrimination of glycosuria in fluorescence spectroscopic data
Rahat Ullah1, Imran Rehan2, Saranjam Khan3
1National Institute of Lasers and Optronics College, Pakistan Institute of Engineering and Applied Sciences, Nilore, Islamabad 45650, Pakistan.
Summary
This study shows fluorescence spectroscopy and random forest machine learning can accurately detect glycosuria (excess sugar in urine) in diabetic patients. The non-invasive method achieved 96% accuracy, offering a promising diagnostic tool.
Area of Science:
- Biomedical diagnostics
- Analytical chemistry
- Machine learning in healthcare
Background:
- Glycosuria, a marker of diabetes, requires accurate and accessible diagnostic methods.
- Current diagnostic approaches may lack convenience or specificity.
- Non-invasive techniques are sought for improved patient outcomes.
Purpose of the Study:
- To evaluate fluorescence spectroscopy combined with machine learning for diagnosing glycosuria.
- To differentiate urine samples from diabetic patients and healthy controls.
- To establish a non-invasive, accurate diagnostic tool for diabetic-related glycosuria.
Main Methods:
- Urine samples analyzed using fluorescence spectroscopy (200-950 nm).
- Random Forest (RF) and K-Nearest Neighbors (KNN) algorithms applied for classification.
- Principal Component Analysis (PCA) used for dimensionality reduction.
Main Results:
- The RF-based model achieved 96% accuracy, 100% specificity, 93% sensitivity, and 100% precision.
- Distinct spectral patterns identified between healthy and diabetic urine samples.
- 10-fold cross-validation confirmed robust model performance.
Conclusions:
- Fluorescence spectroscopy with RF machine learning is a highly accurate method for glycosuria diagnosis.
- The non-invasive approach offers a promising alternative for diabetic patient monitoring.
- This technique could lead to more convenient and precise clinical diagnostics.
