CodeBlanks Logo Icon CodeBlanks
Login Register

What is a metric?

A metric is a numerical value that measures model performance.

What is a metric?

A metric summarises a model's success in a single number. Two separate concepts mustn't be confused:

  • Loss: the function optimised during training (log-loss, MSE).
  • Metric: the measure people use to interpret success (accuracy, F1, MAE).
from sklearn.metrics import get_scorer_names
print(len(get_scorer_names()))          # dozens of ready-made scorers
print("roc_auc" in get_scorer_names())  # True

The metric must reflect the real cost of the problem; the wrong metric leads to a model that looks right but is useless.

Log-loss: quality of probabilities

Log-loss is the mean negative logarithm of the probability given to the correct class. It punishes confident but wrong predictions very hard:

from sklearn.metrics import log_loss
print(round(log_loss([1, 0], [[0.1, 0.9], [0.8, 0.2]]), 3))   # 0.164  good
print(round(log_loss([1, 0], [[0.9, 0.1], [0.2, 0.8]]), 3))   # 1.956  confident and wrong

Accuracy measures only the decision; log-loss measures how well the probabilities are calibrated.

Practice Compare log-loss

For two probability sets that give the same decisions but with different confidence, print the log_loss values with round(..., 3) on separate lines.

Starter code

model.py
from sklearn.metrics import log_loss

y = [1, 0, 1]
emin = [[0.05, 0.95], [0.9, 0.1], [0.2, 0.8]]
kararsiz = [[0.45, 0.55], [0.55, 0.45], [0.4, 0.6]]

# buraya yaz

Sample solution

model.py
from sklearn.metrics import log_loss

y = [1, 0, 1]
emin = [[0.05, 0.95], [0.9, 0.1], [0.2, 0.8]]
kararsiz = [[0.45, 0.55], [0.55, 0.45], [0.4, 0.6]]

print(round(log_loss(y, emin), 3))
print(round(log_loss(y, kararsiz), 3))

Expected output

0.127
0.569

Questions and answers

Review the lesson questions together with their answers. To test yourself, solve the lesson interactively.

1 What does logistic regression minimise in training?

  • Accuracy
  • F1
  • R²
  • Log-loss (cross-entropy) Correct answer

Explanation: Accuracy isn't differentiable; differentiable log-loss is used in training.

2 Compute the log-loss of the probabilities.

Question 
kayip = ____(y_true, olasiliklar)
Answer 
kayip = log_loss(y_true, olasiliklar)

Explanation: log_loss works with predict_proba output.

3 Match the concepts.

Term Meaning
loss Minimised in training
metric Interprets success
log_loss Probability quality
get_scorer_names Ready-made scorers

All lessons in Beginner

1 What is ML?

  1. What is machine learning?
  2. Supervised and unsupervised
  3. Data and model
  4. Training and testing
  5. Feature and label notation
  6. Summary - ML fundamentals

2 NumPy basics

  1. What is NumPy?
  2. Creating arrays
  3. Array properties
  4. Array indexing
  5. Reshaping arrays
  6. Summary - NumPy basics

3 Pandas basics

  1. What is Pandas?
  2. Creating a DataFrame
  3. Column operations
  4. Selecting and filtering data
  5. Viewing first rows
  6. Summary - Pandas basics

4 Data preprocessing

  1. Data cleaning
  2. Missing data
  3. Data types
  4. Label encoding
  5. Data splitting
  6. Summary - Data preprocessing

5 Statistics basics

  1. Mean
  2. Median
  3. Standard deviation
  4. Variance
  5. Correlation
  6. Summary - Statistics

6 Linear algebra

  1. Vectors
  2. Matrices
  3. Matrix multiplication
  4. Transpose
  5. Identity matrix
  6. Summary - Linear algebra

7 Project: Data analysis

  1. Loading data
  2. Cleaning data
  3. Computing statistics
  4. Printing results

8 Regression

  1. What is regression?
  2. Linear regression
  3. Model training
  4. Prediction
  5. Error measurement
  6. Summary - Regression

9 KNN

  1. What is KNN?
  2. The K parameter
  3. Distance measurement
  4. Classification
  5. Model training
  6. Summary - KNN

10 Decision trees

  1. What is a decision tree?
  2. Splitting criterion
  3. Depth
  4. Classification
  5. Overfitting
  6. Summary - Decision trees

11 Classification

  1. What is classification?
  2. Binary classification
  3. Multi-class
  4. Accuracy
  5. Confusion matrix
  6. Summary - Classification

12 Model evaluation

  1. Train-test split
  2. Accuracy
  3. Precision
  4. Recall
  5. F1 score
  6. Summary - Model evaluation

13 Project: Classification

  1. Preparing data
  2. Training model
  3. Prediction
  4. Evaluation

14 Scaling

  1. What is scaling?
  2. StandardScaler
  3. MinMaxScaler
  4. When to use?
  5. Pipeline
  6. Summary - Scaling

15 Feature engineering

  1. What is a feature?
  2. Feature selection
  3. Feature extraction
  4. Categorical features
  5. Numerical features
  6. Summary - Feature engineering

16 Model selection

  1. What is model selection?
  2. Cross-validation
  3. Grid search
  4. Score comparison
  5. Hyperparameter
  6. Summary - Model selection

17 Project: Regression

  1. Preparing data
  2. Training model
  3. Prediction
  4. Error measurement

18 Metrics

  1. What is a metric?
  2. Accuracy
  3. Precision
  4. Recall
  5. F1 score
  6. Summary - Metrics

19 Data visualization

  1. What is visualization?
  2. Scatter plot
  3. Histogram
  4. Line plot
  5. Box plot
  6. Summary - Visualization

20 Beginner final

  1. Challenge 1
  2. Challenge 2
  3. Challenge 3
  4. Challenge 4