Published in

Logistic Regression and Machine Learning

Logistic Regression is helpful and important for Machine Learning Algorithms.

Logistic Regression with the help of an available set of independent variables is useful in predicting the categorical dependent variable. Logistic Regression is a branch of supervised learning techniques of machine learning algorithms. As it is used in solving classification problems. It predicts the output of a categorical dependent variable. In a form of probabilistic values which exist between 0 and 1. Due to its capabilities in delivering the probabilities and with the potential to classify new data using continuous and discrete datasets, it plays an important role in machine learning algorithms.

Sigmoid Function

Sigmoid Function is actually a logistic function as it is an S-shaped curve that takes any real-valued number and which is mapped into a value between 0 and 1, but it is never completely at those limits.

1 / (1 + e^-value)

Here e is the base of the natural logarithms. Value is nothing but the definite value that is supposed to be changed.

Logistic Regressions Assumptions

The variable that is supposed to be independent should not have multi-collinearity.

The variable that is supposed to be dependent should be unconditional in the description.

Logistic Regressions Types

Logistic Regression is classified on the basis of categories.

  • Binomial: The variable that is targeted has only supposed to have only two possible values “o” or “1”.Which can be also stated as, “gain” or “loss”, “win” or “loose”, etc.
  • Multinomial: Variables that are target should have 3 or more possible types and are also supposed not to be any ordered. Like, “disease A” or “disease B” or “disease C”.
  • ordinal: here targeted variables are with ordered categories. For example exam performance can be categorized as: “very poor”, “poor”, “good”, ”very good”. And also every category is given a score like 0,1,2,3.

Mathematical Equations for Logistic Regression.

Lets us now look into mathematical steps to get logistic regression,

  • First, let us consider the equation of a straight line
  • Now we shall divide the above equation by (1-y), as in logistic regression y can be in between o and 1
  • Now we shall consider the logarithm of the equation as the range is supposed to be between -[infinity] to +[infinity].

With this arrive at the equation of logistic regression.

Hence we can say that logistic Regression models are better than any other models because they can classify as well as give probabilities. Hence due to their qualities, they can be extremely useful for AI startups.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store