Understanding Supervised Learning: Regression vs Classification Explained

1. Regression and Classification

Different algorithms can be divided into two categories based on how they “learn” from data to make predictions: supervised learning and unsupervised learning. In supervised learning, you have input variables (x) and output variables (Y) and you use an algorithm to learn the mapping function Y = f(x). The goal is to approximate this function well so that when new input data (x) arrives, the algorithm can predict the corresponding output variable (Y).

Supervised machine‑learning techniques include linear and logistic regression, multi‑class classification, decision trees, and support vector machines. Supervised learning requires that the training data be labeled with the correct answers. For example, a classification algorithm trained on a correctly labeled image dataset of animal species and distinguishing features learns to recognize animals.

Supervised learning problems can be further divided into regression problems and classification problems. Both aim to build a concise model that predicts the value of a dependent attribute from independent attributes. The difference is that regression predicts numeric values, whereas classification predicts categorical labels .

2. Regression

A regression problem occurs when the output variable is a real or continuous value, such as “salary”. Many models can be used, the simplest being linear regression, which attempts to fit the data with the best hyperplane passing through the points.

3. Classification

A classification problem has an output variable that is a category, such as “red” or “blue”, or “disease” versus “no disease”. A classification model tries to draw conclusions from observations. Given one or more inputs, the model attempts to predict the value(s) of one or more outcomes.

For example, predicting whether an email is “spam” or “not spam”. In short, classification either predicts categorical labels or, based on a training set and attribute values (categorical labels), builds a model to classify new data. Many classification models exist, including logistic regression, decision trees, random forests, gradient‑boosted trees, multilayer perceptrons, and naïve Bayes.

References

https://www.geeksforgeeks.org/regression-classification-supervised-machine-learning/

https://machinelearningmastery.com/logistic-regression-for-machine-learning/

https://machinelearningmastery.com/linear-regression-for-machine-learning/