WEBVTT

00:00.080 --> 00:01.000
Welcome back.

00:01.200 --> 00:04.480
The feature correlation heat map is very important.

00:04.480 --> 00:10.680
Once you understand the correlation between the features, you can start pre-processing your data.

00:10.840 --> 00:15.920
So here we need to prepare our data for neural network.

00:15.960 --> 00:25.320
Let's start with x equals to dataset dot drop mpg and y equals to dataset and get the mpg.

00:25.480 --> 00:28.000
Why we dropped the mpg.

00:28.040 --> 00:32.760
Why we removed the mpg column from our data set.

00:32.800 --> 00:37.720
Why we removed this column from the data set?

00:37.920 --> 00:45.720
Because simply, we want to build a model that predicts MPG based on other features.

00:45.960 --> 00:56.280
We therefore need to separate features x the input variables used to make predictions, and target y,

00:56.520 --> 00:59.760
the output variable we want to predict.

00:59.960 --> 01:09.000
So we're going to predict the MPG and based on the x features all columns except MPG.

01:09.280 --> 01:16.640
This includes cylinders, displacement, horsepower, weight, and etc. and target would be only the

01:16.640 --> 01:18.240
mpg column.

01:18.280 --> 01:21.600
Let's print the shape of our data set.

01:21.640 --> 01:26.840
So here we have x dot shape and y dot shape run.

01:26.880 --> 01:27.720
Here we go.

01:27.920 --> 01:31.000
Features shape 392.

01:31.200 --> 01:37.560
With seven we removed one column and the target shape is 392.

01:37.600 --> 01:45.840
So we get the features shape 392 with seven columns and target shape 392.

01:46.040 --> 01:52.360
Your target vector Y has 392 values, one for each data set.