(Summary) Intuition of Machine Learning Algorithms (Need-based approach)

The below shows progression of techniques capable of solving simple to complex tasks in increasing order.

Linear Regression -> Logistic Regression (Non-linear classification method) -> Neural Network (Non-linear, robust method for solving complex tasks)

Logistic regression is a non-linear function more appropriate than linear regression at solving classification problems. It uses sigmoid function, with Linear Regression as input, and is more appropriate for classification problems in that sigmoid function output is between 0 and 1 while Linear Regression outputs between -infinity and + infinity.

To be continued ..