Python for Data Science For Dummies

Chapter 19

Increasing Complexity with Linear and Nonlinear Tricks

In This Chapter

Expanding your feature using polynomials

Regularizing regression

Learning from big data

Using support vector machines

Previous chapters introduced you to some of the simplest, yet effective, machine-learning algorithms, such as linear and logistic regression, Naïve Bayes, and K-Nearest Neighbors (KNN). At this point, you can successfully complete a regression or classification project in data science. This chapter explores even more complex and powerful machine-learning techniques including the following: reasoning on how to enhance your data; controlling the variance of estimates by regularization; and managing to learn from big data by breaking it into manageable chunks.

This chapter also introduces you to the support vector machine (SVM), a powerful family of algorithms for classification and regression. SVMs are able to perform the most difficult data problems and are a perfect substitute for neural networks such as the multilayer perceptron, which isn’t currently present in the Scikit-learn package ...