# Machine Learning Introduction

From Software Engineers Wiki

## Contents |

## Basic

What is machine learning: Develop learning algorithm. Using training set, machine finds the hypothesis that can predict target out of features.

Basic terms:

- x(i): feature, or input
- y(i): target, or output
- (x(i), y(i)): training example
- m: number of training example
- data set of training example: training set
- i: index into the training set
- h: hypothesis
- theta: parameter, or weight

Type of problem:

- regression problem: target variable to predict is continuous
- classification problem: target can take on a number of discrete values

Learning style:

- supervised learning
- unsupervisied learning
- semi-supervised learning

Regression (statistical machine learning) algorithms:

- linear regression
- ordinary least squares regression (OLSR)
- logistic regression
- stepwise regression
- multivariate adaptive regression splines (MARS)
- locally estimated scatterplot smoothing (LOESS)