선형 회귀 학습하기

Linear regression fits a straight-line relationship between your features and a numeric target. It is the classic first model for a reason. 📈

The estimator lives in the linear_model module. You import it once and reuse it for any regression task you meet.

Build a fresh model by calling the class. This blank estimator holds default settings and has not seen any data yet.

Now teach it with your features and target. During fit, the model finds the best line through your training points.

Each feature gets a coefficient stored in coef_. It tells you how much the prediction moves when that feature rises by one.

The intercept is the prediction when every feature is zero. It anchors the line and is stored in intercept_ after fitting.

Hand new feature rows to predict and you get numeric estimates back, one per row, computed straight from the fitted line.

Under the hood, each prediction is just features times coefficients plus the intercept. Simple math, surprisingly powerful results.

The same call handles a single feature or dozens. With many inputs it fits a hyperplane, but your code stays exactly the same.

Linear regression assumes a roughly straight relationship. If the pattern curves sharply, this model will underfit and miss it.

Import, create, fit, predict: four steps and you have a working regressor. This same rhythm repeats for every model you learn next.

Which attribute holds the per-feature weights after fitting?

You imported, created, fit, and predicted with a linear regression. Its coef_ and intercept_ even reveal what it learned. 🚀