분류 모델

A model whose prediction is a class. For example, the following are all classification models:¹

A model that predicts an input sentence’s language (French? Spanish? Italian?).
A model that predicts tree species (Maple? Oak? Baobab?).
A model that predicts the positive or negative class for a particular medical condition.

Types of classification model

Binary classification
Multi-class classification

Metrics

Accuracy:

Use as a rough indicator of model training progress/convergence for balanced datasets.
For model performance, use only in combination with other metrics.
Avoid for imbalanced datasets. Consider using another metric.

Recall:

Use when false negative are more expensive than false positives.

False positive rate:

Use when false positives are more expensive than false negatives.

Precision:

Use when it’s very important for positive predictions to be accurate.

Footnotes

developers.google.com/machine-learning/glossary#classification_model

Introduction to machine learning problem framing 이 글을 인용: “"Classify an email as spam or not spam."”
Classification threshold “In a binary classification, a number between 0 and 1 that converts the raw output of a logistic regression mode…”
Introduction to Machine Learning 관련 절: Definition of [[Machine learning|ML]]
ML crash course - Linear regression 관련 절: Convergence and convex functions
지도 학습 “Training a model from features and their corresponding labels.”
로지스틱 회귀 관련 절: Explanation
Positive class (machine learning) “In binary classification, the class you are testing for.”
Self-training (machine learning) “Use supervised learning to train a model on the labeled examples.”
Area under the ROC curve “A number between 0.0 and 1.0 representing a binary classification model's ability to separate positive classes …”
k-DT: A Multi-Tree Learning Method 이 글을 인용: “The purpose of this technique is to overcome small errors that appear in individual classifiers.”
Negative class (machine learning) “In binary classification, one class is termed positive and the other is termed negative.”
Downsampling (machine learning) “For example, in a class-imbalanced dataset, models tend to learn a lot about the majority class and not enough …”
ROC curve “The shape of an ROC curve suggests a binary classification model's ability to separate positive classes from ne…”
Minority class (machine learning) “For example, given a dataset containing 99% negative labels and 1% positive labels, the positive labels are the…”
Generalization (machine learning) “To encourage generalization, regularization helps a model train less exactly to the peculiarities of the data i…”
선형 모델 “A model that assigns one weight per feature to make predictions.”
Iteration (machine learning) “A backward pass (backpropagation) to adjust the model's parameters based on the loss and the learning rate.”
기계학습 관련 절: Generalization and OOD
Parameter (machine learning) “For example, learning rate is a hyperparameter.”
Learning rate “If you set the learning rate too high, gradient descent often has trouble reaching convergence.”

Types of classification model

Metrics

See also

Footnotes