# TIL Cost Functions permalink

I’m excited to have started Andrew Ng’s *Machine Learning* class oun Coursera over the weekend!
I hope to learn how to apply machine learning to my decision-making process to make better, more informed decisions.

I’ve only gone through the first few lectures which have reacquainted me with basic machine learning concepts:

- supervised vs unsupervised learning
- regression, classification, and clustering
- parameters, features, variables and labels
- linear algebra and matrix multiplication

But the biggest thing I’ve learned so far is what a *cost function* is and why it is important.
A cost function is an algorithm that measures the accuracy of the algorithm in your model (otherwise know as your *hypothesis*).
This cost function is an important part of the learning process because it helps you quickly identify the optimal algorithm based on your training data.

The class went through building a linear regression algorithm that used a *mean squared error* cost function to evaluate how using different function parameters affected the accuracy of the algorithm.