Supervised Learning—Classification Using Logistic Regression
Wei-Meng Lee · 2019
This chapter helps the coders to learn a supervised machine learning algorithm—logistic regression. The main problem with linear regression is that the predicted value does not always fall within the expected range. The chapter shows how logistic regression solves this problem. It discusses the details of the logistic regression algorithm —odds. When the coders apply the natural logarithm function to the odds, they get the logit function. The chapter explains how to transform logit function into a sigmoid function. Scikit-learn ships with the Breast Cancer Wisconsin (Diagnostic) Data Set contains 30 features, and they are computed from a digitized image of a fine needle aspirate (FNA) of a breast mass. The chapter discusses some of the metrics that are useful in determining the effectiveness of a machine learning algorithm. In addition, the coders learned about what a Receiver Operating Characteristic curve is, how to plot it, and how to calculate the area under the curve.