WEBVTT 00:00.080 --> 00:00.960 Welcome back. 00:01.200 --> 00:05.320 We calculated the metrics MSE and R squared score. 00:05.560 --> 00:14.080 Now let's create a scatter plot with regression line to visualize how well your model fits the test 00:14.080 --> 00:14.680 data. 00:14.720 --> 00:20.240 Okay, so at first let me start with PLT dot figure. 00:20.360 --> 00:22.720 Figure size is 8.5. 00:22.760 --> 00:24.880 Let me create a scatter. 00:25.080 --> 00:29.400 So eight by five eight inches wide and five inches tall. 00:29.440 --> 00:40.920 Plot scatter x test y test color blue and label actual and another scatter x test y predict. 00:40.960 --> 00:51.600 You see guys, I am using Xtest input feature and y test which is the actual target values. 00:51.840 --> 00:54.240 Then I used x test. 00:54.240 --> 00:59.080 And why predict y predicts for model's predictions. 00:59.400 --> 01:03.920 Okay, so Plitt predicted values as red x marks. 01:03.920 --> 01:06.440 So here we have red. 01:06.480 --> 01:08.960 Let me make a note. 01:09.080 --> 01:16.040 This plots the predicted values as red x marks and here as blue dots. 01:16.280 --> 01:19.480 So the actual values are in blue. 01:19.720 --> 01:23.280 The predicted values are in red. 01:23.320 --> 01:23.720 Okay. 01:23.760 --> 01:34.000 So we're going to display a graph that gives us the red dots as predicted and the blue dots as actual. 01:34.000 --> 01:45.280 And we're going to see the difference and how well the predicted is similar or is closer or is far or 01:45.360 --> 01:48.760 different from the actual data. 01:48.920 --> 01:59.240 Then plotting start plotting with X, test y predict color red line width two and label fitted line. 01:59.280 --> 02:07.280 This connects predicted values with a red line and shows the regression line curve that your model learned. 02:07.320 --> 02:12.120 Now let's add the title and the label for our graph. 02:12.280 --> 02:13.360 Plot title. 02:13.360 --> 02:19.880 Actual versus predicted prices x label size square 1000 square meter. 02:19.920 --> 02:29.560 Plot the y label y axis as the price and plot legend grid true and show the plot. 02:29.600 --> 02:32.120 Let's run and here we go. 02:32.240 --> 02:35.320 This is our amazing result. 02:35.360 --> 02:37.600 The red line is the predicted. 02:37.600 --> 02:39.320 You can see the legend here. 02:39.360 --> 02:50.360 Red line is the predicted actual price is the blue dots as the points evenly distributed around the 02:50.360 --> 02:50.880 line. 02:51.200 --> 02:53.840 Also clear linear relationship. 02:54.000 --> 03:03.090 And as the blue dots close to the red line as you get a good fit and a good result. 03:03.090 --> 03:10.730 So as those line, close those blue dots closer cut. 03:10.770 --> 03:12.050 We have this legend. 03:12.050 --> 03:18.530 Actual dots are in blue that predicted in in X and the fitted line here. 03:18.530 --> 03:24.730 So as we get closer to the red line as we get a good fit. 03:24.730 --> 03:33.290 So the blue lines actual close to red line, which is the predicted points evenly distributed around 03:33.290 --> 03:36.690 the line and clear linear relationship. 03:36.690 --> 03:38.490 This is a good result. 03:38.490 --> 03:43.370 So here I want from you to understand this note. 03:43.370 --> 03:48.050 For the good results you should get for good results. 03:48.050 --> 03:54.610 As I told you, you should get the blue dots close to the red line. 03:54.610 --> 04:03.210 So you can see those blue dots are close to the red line points evenly distributed around the line. 04:03.210 --> 04:09.250 Also we have evenly distributed points clear linear relationship. 04:09.250 --> 04:13.050 So we can see that we are following linear line. 04:13.050 --> 04:14.610 So this is the straight line. 04:14.610 --> 04:19.610 If we get blue dots far from red line. 04:19.650 --> 04:30.450 Curved pattern in residuals points clustered in certain areas you get poor fit and bad results. 04:30.650 --> 04:40.610 So you should get those symptoms, those indications, the blue dots close to red line points evenly 04:40.610 --> 04:48.810 distributed around the line, and the clear linear relationship in order to get a good result for your 04:49.050 --> 04:49.610 model.