Fix inconsistent model training approach in trian_and_predict function

[sohampirale]

Summary

This PR fixes an inconsistency in the trian_and_predict function where models were not being handled according to their hyperparameter configurations.

Fixes : #71

---

Solution

Updated the conditional logic to properly align model usage with hyperparameter availability:

• IF block (simple fitting): Models without hyperparameters in config.py

  • "Linear Regression"
  • "Random Forest Regressor"

• ELSE block (GridSearchCV): Models with hyperparameters in config.py

  • "Ridge Regression", "Lasso Regression", "Elastic Net", "Bayesian Ridge", "SGD Regressor", "Support Vector Regressor"

---

Files Changed

• app/services/gaze_tracker.py: Updated the conditional logic in trian_and_predict function

---

Screenshots (tested with all 3 models)

• Elastic Net

• Support Vector Regressor

• Random Forest Regressor