Model Evaluate

Category: Mega-Polis → Analysis → Analysis Data Tools
Node ID: SvMegapolisModelEvaluate
Tooltip: Model Evaluate
Dependencies: sklearn

Functionality

Evaluates regression predictions using scikit-learn metrics.

Given: - a model object (not directly used for scoring here), - prediction values, - and ground-truth y values,

the node computes:

  • R² score (sklearn.metrics.r2_score)
  • RMSE (sklearn.metrics.mean_squared_error(..., squared=False))

and is intended to output both scores.

Inputs

Socket Type Description
Model SvStringsSocket A fitted model object (expects list-wrapped: model = self.model[0]). Currently not used in the calculations.
Predictions SvStringsSocket Predicted values (expects list-wrapped: predictions = self.predictions[0]).
y SvStringsSocket Ground-truth values y_test (the code uses y_test = self.y without indexing).

All three inputs must be linked for the node to execute.

Parameters

This node has no exposed UI parameters.

Outputs

Socket Type Description
r2 SvStringsSocket Intended to output the computed R² score.
rmse SvStringsSocket Intended to output the computed RMSE.

Example

Evaluate a regression model

  1. Use your pipeline to produce:
    • a fitted Model
    • Predictions from that model
    • ground-truth y values
  2. Connect:
    • Model → Model
    • Predictions → Predictions
    • y_test → y
  3. Read outputs:
    • r2 for goodness-of-fit
    • rmse for error magnitude in the target units