- Support Vector Machines → Developed as a classification method within the framework of statistical learning theory.
- Generalized to handle regressions, both linear and nonlinear, known as Support Vector Regressions.
- This method looks to do convex optimizations, and geometrically is based on maximum margin hyperplane classifiers.
- The use of a kernel allows SVM to control non-linear architecture and parameters.
The basis of any machine learning model follows the structure shown below (in its simplest form), using the syntax as defined in scikit-learn.
from sklearn.module import Model
model = Model()
model.fit(X, y)
predictions = model.predict(X_new)
print(predictions)We’ll look at how we use this syntax to have scikit-learn train a classifier model using K-Nearest Neighbors algorithm.
from sklearn.neighbors import KNeighborsClassifier
# We use .values to convert X and y to arrays
X = churn_df[["total_day_charge", "total_eve_charge"]].values
y = churn_df["churn"].values
knn = KNeighborsClassifier(n_neighbors=15)
knn.fit(X,y)
X_new = np.arrayMachine learning is broken down to a few types, most notably: supervised and unsupervised. Focusing on supervised learning, where you train machine learning models with “labeled” data for it to learn from, we have two specific types:
- Classification → Target variable consists of categories
- Regression → Target variable is continuous
Classifying labels of unseen data
- Build a model
- Model learns from the labeled data we pass to it
- Pass unlabeled data to the model as input
- Model predicts the labels of the unseen data
- Labeled data = training data
Accuracy = # of correct predictions divided by total observations