Machine Learning Crash Course: Part 1

In other words, by giving our algorithm examples of apples and oranges to learn from, it can generalize its experience to images of apples and oranges that it has never encountered before. This type of machine learning--drawing lines to separate data--is just one subfield of machine learning, called classification. For example, square footage is a good predictor of house prices, so our algorithm should give square footage a lot of consideration by increasing the coefficient associated with square footage. In our example of predicting house prices based on square footage, since we're only considering one variable our model only needs one input feature, or just one x: Now the question becomes: How does a machine learning algorithm choose c2 and c1 so that the line best predicts house prices?