Web AI: Varicose veins prediction

Click on this world map at your location. Then, fill in the fields with your age, gender, and body mass index, and press the "Predict" button. The chance of getting varicose veins will be displayed as a percentage (%).

All calculations will be performed on the client side in the browser, ensuring privacy with no data transfer to third-party servers.

Color Values Grace
Version: 2.0
Deployment Date (Update): 2025-01-02 09:32:03
Author: Dr. Yury Rusinovich
Email: yury.rusinovich@gmail.com
Model Training Platform: MLHS Playground
AI/ML Library: TensorFlow.js
References:
Research paper 1 & Data source
Research paper 2
Source of the map

Performance and Training Settings:

Test Loss: 0.49169594049453735

Validation Loss: 0.9123176336288452

Training Loss: 0.876839816570282

Test MAE: 0.5590166449546814

Validation MAE: 0.8001748323440552

Training MAE: 0.6520715355873108

R2 score: 0.362848530381088

Data Counts (n): 81

Outcome: Varicose % (Outcome): Mean = 23.455, SD = 11.972

Predictors (n): Gravity of the region (mGal): Mean = 15.481, SD = 14.356, p-Value = 0.001; Age in years: Mean = 41.228, SD = 10.859, p-Value = 0; Gender (0=male, 100=female): Mean = 60.06, SD = 27.934, p-Value = 0.657; Body mass index ( kg/m²): Mean = 26.058, SD = 1.491, p-Value = 0.203

L2 Legularization: 0.01

Number of Epochs: 406

Initial Learning Rate: 0.001

Batch Size: 32

Data preprocessing: Z-score scaling



Add a Comment:

This app has been used 0 times for predictions.

This machine learning model could be downloaded and retrained by adding new data here