Continuing from the previous ①, we will divide the training data and test data a little more practically.

First, prepare the data.

Now there are two variables x. Thus, in machine learning, when there are two or more variables x We will combine them into one and treat them as uppercase $ \ mathbf {X} $.

Next, divide it into training data and test data.

Training data test data Then, we will create a model formula from the training data. If this is calculated by the method of least squares, it can be calculated as follows.

And we will apply this to the test data.

From the above, for y_test of the correct answer data I found that the y_pred I expected was roughly correct.

If there are two or more $ x $ for $ y $ Basically, think of training data and test data as above.

In actual machine learning, for data like this one y is the rent, x1 and x2 are the constituent elements (station walk, age, etc.) We are analyzing the data.

Also, the previous article ① and this article ② Both of them use a method called linear regression. I hope to post again in the near future.

Recommended Posts