WEBVTT 00:00.080 --> 00:01.080 Welcome back. 00:01.120 --> 00:09.120 We used numpy library to generate random numbers and create a linear relationship between size and price. 00:09.320 --> 00:16.840 Then we use the data frames from the pandas library to create this table size and the price. 00:17.120 --> 00:28.120 Also, we use the Matplot library to visualize and display the created and generated data into this 00:28.160 --> 00:28.880 chart. 00:29.040 --> 00:35.600 So this chart displays only the first ten results of the data that we created. 00:35.600 --> 00:38.120 Also, this table do the same job. 00:38.280 --> 00:38.600 This. 00:38.600 --> 00:45.160 Thanks for the head function that gets all the first ten rows. 00:45.200 --> 00:55.160 Now let's deep dive into machine learning how to use those data into our model to train and prepare 00:55.440 --> 00:56.560 the model. 00:56.600 --> 00:59.600 The first step is to prepare the data. 00:59.600 --> 01:02.310 We split data into two things. 01:02.310 --> 01:05.910 Two parts training sets to train the model. 01:05.950 --> 01:09.150 Testing set to evaluate performance. 01:09.270 --> 01:13.630 We use train test split for scikit learn. 01:13.670 --> 01:14.990 Let me show you guys. 01:15.030 --> 01:19.630 First we have created two variables x and y. 01:19.950 --> 01:22.190 X equals to data. 01:22.230 --> 01:22.910 Data. 01:23.150 --> 01:26.990 Did you remember the data which is the data frame that we created before. 01:26.990 --> 01:34.030 So if you scroll up we see that we stored all the data inside the data variable. 01:34.030 --> 01:36.110 In order to access the size. 01:36.150 --> 01:45.430 We access it using this angled bracket and y equals to data and accessing the price and square in thousands 01:45.430 --> 01:46.670 of dollars. 01:46.710 --> 01:56.070 Now create those random numbers and random variables again guys X is the feature, Y is the target. 01:56.110 --> 01:58.710 Here we have the split. 01:58.790 --> 02:01.780 We have to split the data at first. 02:01.820 --> 02:12.140 Xtrain Xtest Ytrain y test equals to train test split x y test size and Randomstate. 02:12.300 --> 02:14.500 What does this means? 02:14.660 --> 02:25.780 Train test split is a function from SK learn dot model selection that splits your data set into training, 02:25.820 --> 02:33.620 test set and testing set training set X train and Y train used to train the model. 02:33.740 --> 02:42.540 Testing set X test and Y test are used for for evaluating the model's performance. 02:42.580 --> 02:52.180 The parameters are x, y test size and random set x and y that do constants, the data frame and the 02:52.180 --> 02:52.940 series. 02:52.980 --> 03:04.360 The feature and the target that test size 0.2 is 20% of the data will be for testing and 80% for training. 03:04.360 --> 03:06.760 So let me write this node down. 03:06.960 --> 03:17.480 0.2 equals to 20% of the data will be for testing and 80% for training. 03:17.600 --> 03:23.880 Random state equals to 42 ensures that the split is reproducible. 03:24.120 --> 03:27.960 You get the same split every time you run it. 03:28.000 --> 03:29.880 Okay, it's very simple. 03:29.920 --> 03:34.040 Now checking the numbers of the sample training. 03:34.040 --> 03:36.680 Sample x train dot shape. 03:37.000 --> 03:42.200 Train shape zero and testing the samples Xtest dot shape zero. 03:42.320 --> 03:44.360 The X train shape. 03:44.360 --> 03:45.960 Here we are training. 03:45.960 --> 03:50.520 The model gives the number of rows samples in the training set. 03:50.520 --> 03:56.280 So gives the number of rows in the training set. 03:56.280 --> 04:00.960 And this line gives the number of samples in the test set. 04:01.310 --> 04:08.190 This helps confirm that the split worked and as expected, let's run and here we go. 04:08.230 --> 04:20.070 Training sample 80% and testing sample is 20% okay so this is how we split the data into uh into testing 04:20.070 --> 04:22.270 and training sets. 04:22.310 --> 04:31.590 Again as a quick summary, what we've done before X is the feature, the house size, the y axis and 04:31.590 --> 04:37.910 the y variable is the target which is the price split to 20%. 04:37.910 --> 04:46.910 So the test size would be 20% and 80% for training and random state equals to 42. 04:46.950 --> 04:53.990 Print the number of samples in each set and every time get the same split every time you run it. 04:54.150 --> 04:58.350 Okay, this is how we split the data into testing and to train.