4.1 Introduction

There are multiple Machine Learning problems that require predicting a categorical value, such as “true or false”, “churn or not churn”, “fraud or not fraud”, “high risk, low risk or medium risk”, etc. These are called Classification problems, and there can be multiple categories (or classes) to predict.

Logistic regression is a Supervised learning Machine Learning technique that can be used to solve classification problems. These problems can also be solved with other Machine Learning methods, such as models, ensembles or deepnets. We explain these methods in Chapter 1 , Chapter 2 and Chapter 5 respectively. The main difference is that logistic regression assumes your input Fields can be mapped to predict your Objective Field following linear patterns. For this reason, logistic regressions work better in those cases for which the problem can be linearly solved.

For each class of the objective field, logistic regression computes a probability modeled as a logistic function value, whose argument is a linear combination of the field values. See section 4.2 for more details on the logistic regression formula.

This chapter contains comprehensive description of BigML’s logistic regressions including how they can be created with 1-click (section 4.3 ), all configuration options available (section 4.4 ), and the different visualizations provided by BigML (section 4.5 ). See section 4.6 for an explanation of how logistic regressions can be used to make predictions. You can also export your logistic regressions in different formats to make local predictions faster at no cost (section 4.6 ). The process to evaluate your logistic regressions’ predictive performance in BigML is explained in a different chapter (Chapter 7 ).

In BigML, the third tab of the main menu on the Dashboard allows you to access all of your available Supervised learning models. Select Logistic Regressions from the drop-down menu (Figure 4.1 ), you will reach the logistic regression list view.

The logistic regression list view (Figure 4.2 ) lists all your available logistic regressions. For each logistic regression, the view shows the link to the Dataset used to create it, its Name, Objective (Objective Field name), Age (time elapsed since it was created), Size, and number of evaluations, predictions, and batch predictions that have been created using that logistic regression. The search menu option in the top right corner of the logistic regression list view allows you to search your logistic regressions by name.

By default, when you first create an account at BigML, or every time that you start a new Project, your list of logistic regressions will be empty. (See Figure 4.3 .)

Finally, in Figure 4.4 you can see the icon used to represent a logistic regression.