Supervised Learning
Given an algorithm that requires some data sets already have the correct answer. For example, given the price data set. Supervised learning is also called regression.
Example: House price prediction, cancer prediction
Unsupervised Learning
The sample set is not labeled, and a set of unlabeled data is divided into multiple clusters
Example: Organizing computer clusters, social network analysis
Cocktail Party Issues
Extracting effective information from background noise.
[W,s,v]=svd ((Repmat (sum (x.*x,1), size (x,1), 1). *x) *x ');
linear regression
For example, prices are as follows:
$x _{1}^{(i)}$ represents the living area of the first house, $x _{2}^{(i)}$ represents the number of bedrooms in the first house, so X is a two-dimensional vector.
Define hypothetical functions: $h _{\theta} (x) $=$\theta_{0}$+$\theta_{1}$ $x _{1}$+$\theta_{2}$ $x _{2}$
$h (x) =\sum_{i=0}^{n}\theta_{i}x_{i}$ = $\theta ^{t}x$
The next question is how to solve the $\theta$ value
Define the following loss function to indicate the proximity of $h (X^{i}) $ to $y^{i}$
$J (\theta) =\frac{1}{2}\sum_{i=1}^{m} (H_{\theta} (x^{(i)})-y^{{i}) ^{2}$
LMS Algorithm
Supervised learning and unsupervised learning