Basic Logistic Regression for MR

Logistic Regression Classifier in Python - Basic Introduction

In logistic regression... basically, you are performing linear regression but applying a sigmoid function for the outcome.

Sigmoid / Logistic Function

𝑝=1/1+𝑒−𝑦

Properties of Logistic Regression

The dependent variable follows a Bernoulli Distribution Estimation is maximum likelihood estimation (MLE)

Advantages

Straight forward, easy to implement, doesn't require high compute power, easy to interpret, used widely. Doesn't require feature scaling and provides a probability score for observations.

Disadvantages

Not able to handle a large number of category features/variables. Vulnerable to overfitting.

Data

Is the candy chocolate? Let's find out because... yum

(This intro was built on a candy-data.csv dataset in DataScience folder also found at

https://github.com/fivethirtyeight/data/blob/master/candy-power-ranking/candy-data.csv )