Random Forest Regressor Model Predicted Well

A project log for Open Muscle Finger Tracking Sensor

Forearm Muscle-Based Finger Tracking Device

TURFPTAxTURFPTAx 03/26/2023 at 22:080 Comments

The trained model had a Mean Average Error (MAE) value of approximately 25, equating to a 12% error rate. Despite the challenges, the model was able to make real-time predictions, demonstrating its potential for a significant impact on the prosthetic industry.

Predicted Vs Actual from finger 1 or LASK1. The index finger was the most predicted pattern from the open muscle band. Here you see the Actual data from LASK1 in blue and the predicted output of the LASK1 from the model in orange. These are just preliminary data. Everything is published live on the gGitHub The model is 1.7GB so we are not providing the model but all the code is available to train the model with this exact dataset on the gGitHub From attachment predictions_vs_actuals2.csv

GitHub Link for Open Muscle: