WEBVTT 00:00.080 --> 00:01.080 Welcome back. 00:01.120 --> 00:04.800 We trained our model and get the learned equation. 00:04.800 --> 00:09.040 Price equals to 2.9 times size plus 2.7. 00:09.320 --> 00:17.280 This is the equation that it is concluded by the model when is being trained. 00:17.320 --> 00:18.120 Go up. 00:18.120 --> 00:21.080 If you remember we splitted the data. 00:21.120 --> 00:29.320 20% of the data will be used for testing and 80% of the data will be used for training. 00:29.360 --> 00:32.000 We used this 80% for training. 00:32.000 --> 00:37.080 Now we're going to use this 20% of the data for testing. 00:37.080 --> 00:41.120 So step number three make predictions on test data. 00:41.320 --> 00:54.280 So here let me start testing y pred equals to model dot predict x test model dot predict is uses the 00:54.280 --> 00:59.080 trained model to make predictions on the test data. 00:59.120 --> 01:03.480 Set X test contains the input features for testing. 01:03.680 --> 01:07.960 X-test contains the input features for testing. 01:08.200 --> 01:13.000 Why predict stores the model's predicted values? 01:13.160 --> 01:16.400 If we run, everything will work fine. 01:16.440 --> 01:20.320 Now let's calculate the evaluation metrics. 01:20.320 --> 01:26.880 Start with MSE equals to mean squared error y. 01:27.160 --> 01:32.880 Test y predict the MSE mean squared error. 01:32.920 --> 01:39.640 Measures the average squared difference between actual and predicted values. 01:39.840 --> 01:43.000 Lower MSE better performance. 01:43.000 --> 01:49.240 So when you get lower MSE, your uh, you're getting better performance. 01:49.280 --> 01:58.520 Now let's calculate another thing that we're going to use for evaluating our metrics, which is the 01:58.520 --> 02:08.160 r squared score R squared score R squared score measures how well the model explains the variance in 02:08.160 --> 02:08.720 the data. 02:08.760 --> 02:12.680 Higher R square equals to better performance. 02:12.840 --> 02:16.320 The range is from minus infinity to one. 02:16.520 --> 02:20.680 So r square equals to r squared score. 02:20.880 --> 02:25.520 What data we need to pass the y test and the y prediction. 02:25.520 --> 02:31.640 So we need to set the difference between predicted and the tested once. 02:31.640 --> 02:35.040 So the actual and the predicted. 02:35.200 --> 02:42.440 Once you see guys how we compared how we pass the data for the testing and prediction. 02:42.480 --> 02:48.600 How to set the difference by MSE and how to calculate a measure. 02:48.640 --> 02:52.360 How the model explains the variance in data. 02:52.360 --> 02:59.200 So MSE we need to get lower MSE and we need to get higher R squared. 02:59.240 --> 03:06.970 Don't worry we're going to clarify everything in the in the chart in the next couple of minutes here. 03:06.970 --> 03:12.930 We need to print mean squared error MSE and r squared score. 03:12.970 --> 03:15.130 Let's run our application. 03:15.530 --> 03:17.450 Here we get the error. 03:17.490 --> 03:21.530 We need uh the name model is not defined. 03:21.530 --> 03:24.090 So let me go up. 03:24.130 --> 03:26.530 And because I started the session. 03:26.650 --> 03:31.730 So I need to run this run this, run this. 03:37.970 --> 03:40.170 And then run this okay. 03:40.490 --> 03:44.130 So we get mean squared error. 03:44.170 --> 03:48.490 The MSE is 0.163. 03:48.530 --> 03:54.170 For r squared score is 0.9507. 03:54.210 --> 04:03.850 In order to see if you get a good result or not it depends on the data and data scale. 04:03.850 --> 04:08.770 So in order to determine that if this MSA is good or not. 04:08.810 --> 04:15.530 We go up and measure how many data set we have. 04:15.850 --> 04:21.370 So in our data set we have around 100. 04:21.370 --> 04:29.930 So 100 times 0.16 is good equals to 16% error. 04:30.050 --> 04:34.770 Cut to see if MSA is good or not. 04:34.930 --> 04:39.690 We need to check the target values and your data scale. 04:39.690 --> 04:47.650 If we go up, we have 100 rows of data that are going to be tested. 04:47.650 --> 04:51.530 So here we have 100 rows. 04:51.530 --> 04:54.730 So generating 100 random numbers. 04:54.730 --> 04:56.170 This is our data. 04:56.170 --> 04:59.130 So our data is being displayed. 04:59.410 --> 05:03.330 And here we are displaying only the first 110. 05:03.570 --> 05:10.330 If I need to get the whole data just run the application and you can see that it is 99. 05:10.530 --> 05:11.490 Starting from zero. 05:11.490 --> 05:12.930 So it's 100. 05:12.970 --> 05:17.690 Okay, so guys we are testing our data on 100. 05:17.730 --> 05:25.770 And we get this amazing result which is 0.6 which is perfect and outstanding. 05:25.770 --> 05:32.450 So for from 0 to 100 we get 0.6 which is good. 05:32.490 --> 05:40.410 But I want from you to focus with me 0.63 for for data range. 05:40.650 --> 05:45.330 For data range between 0 and 1. 05:45.530 --> 05:54.650 It's quite quite poor 0.63 for for data between 0 and 10. 05:54.810 --> 05:55.650 Excellent. 05:55.650 --> 05:58.930 Ranging between 0 and 100 is outstanding. 05:58.930 --> 06:00.290 So this is our case. 06:00.290 --> 06:02.850 It's very good okay guys. 06:02.850 --> 06:11.530 So this is a very important thing you should learn which is the metrics the mean and the r squared.