Model Evaluation is the process of measuring how well a Machine Learning model performs on unseen data.<\/p>\n\n\n\n
After training a model, we must check its performance to ensure it generalizes well and does not just memorize training data.<\/p>\n\n\n\n
Model evaluation helps:<\/p>\n\n\n\n
Measure accuracy
Detect overfitting and underfitting
Compare different models
Improve model performance
Ensure reliability before deployment<\/p>\n\n\n\n
Without evaluation, we cannot trust predictions.<\/p>\n\n\n\n
Data is usually divided into:<\/p>\n\n\n\n
Training Set \u2192 Used to train the model
Testing Set \u2192 Used to evaluate performance<\/p>\n\n\n\n
This ensures the model is tested on unseen data.<\/p>\n\n\n\n
Common split:<\/p>\n\n\n\n
70\u201380% Training
20\u201330% Testing<\/p>\n\n\n\n
Regression models predict continuous values.<\/p>\n\n\n\n
Average of absolute differences between actual and predicted values.<\/p>\n\n\n\n
Lower MAE means better model.<\/p>\n\n\n\n
Average of squared differences.<\/p>\n\n\n\n
Penalizes larger errors more than MAE.<\/p>\n\n\n\n
Square root of MSE.<\/p>\n\n\n\n
Easier to interpret because it is in original units.<\/p>\n\n\n\n
Measures how well the model explains variance in data.<\/p>\n\n\n\n
Value range:<\/p>\n\n\n\n
0 to 1<\/p>\n\n\n\n
Closer to 1 means better fit.<\/p>\n\n\n\n
Classification models predict categories.<\/p>\n\n\n\n
Percentage of correct predictions.<\/p>\n\n\n\n
Accuracy = Correct Predictions \/ Total Predictions<\/p>\n\n\n\n
Best used when data is balanced.<\/p>\n\n\n\n
Shows:<\/p>\n\n\n\n
True Positive (TP)
True Negative (TN)
False Positive (FP)
False Negative (FN)<\/p>\n\n\n\n
Helps analyze model errors.<\/p>\n\n\n\n
Precision = TP \/ (TP + FP)<\/p>\n\n\n\n
Measures how many predicted positives are correct.<\/p>\n\n\n\n
Important when false positives are costly.<\/p>\n\n\n\n
Recall = TP \/ (TP + FN)<\/p>\n\n\n\n
Measures how many actual positives are identified.<\/p>\n\n\n\n
Important when false negatives are costly.<\/p>\n\n\n\n
Harmonic mean of precision and recall.<\/p>\n\n\n\n
Useful when dataset is imbalanced.<\/p>\n\n\n\n
Instead of using one train-test split, cross-validation divides data into multiple folds.<\/p>\n\n\n\n
Example:<\/p>\n\n\n\n
5-Fold Cross Validation:<\/p>\n\n\n\n
Data split into 5 parts
Train on 4 parts
Test on 1 part
Repeat 5 times<\/p>\n\n\n\n
Final score is average of all folds.<\/p>\n\n\n\n
Benefits:<\/p>\n\n\n\n
More reliable evaluation
Reduces bias
Better performance estimation<\/p>\n\n\n\n
Overfitting:<\/p>\n\n\n\n
Underfitting:<\/p>\n\n\n\n
Good model:<\/p>\n\n\n\n
Regression evaluation:<\/p>\n\n\n\n
from sklearn.metrics import mean_squared_error, r2_scoremse = mean_squared_error(y_test, predictions)
r2 = r2_score(y_test, predictions)print(\"MSE:\", mse)
print(\"R2 Score:\", r2)<\/pre>\n\n\n\nClassification evaluation:<\/p>\n\n\n\n
from sklearn.metrics import accuracy_score, classification_reportprint(\"Accuracy:\", accuracy_score(y_test, predictions))
print(classification_report(y_test, predictions))<\/pre>\n\n\n\nChoosing the Right Metric<\/h2>\n\n\n\n
Use MAE, MSE, RMSE for regression
Use Accuracy for balanced classification
Use Precision\/Recall for imbalanced data
Use F1-score when both precision and recall matter<\/p>\n\n\n\nKey Takeaway<\/h2>\n\n\n\n
Model Evaluation measures how well a Machine Learning model performs on unseen data.<\/p>\n\n\n\n
Using proper metrics ensures the model is accurate, reliable, and ready for real-world applications.<\/p>\n\n\n