WEBVTT 00:00.080 --> 00:01.120 Welcome back. 00:01.120 --> 00:02.600 Now scroll down. 00:02.640 --> 00:05.880 We have a data set simulated house prices. 00:05.880 --> 00:13.160 We're going to create a simple data set where the size in square feet which is a feature on x axis, 00:13.440 --> 00:18.120 and the price in thousands, which is the target on y axis. 00:18.120 --> 00:21.480 Did you remember the feature and the target terminology? 00:21.520 --> 00:31.120 Please go back to the previous lessons and in the in the ML introduction section for more information. 00:31.320 --> 00:34.760 This is a supervised regression problem. 00:34.760 --> 00:44.520 Supervised means predicting a continuous numerical value based on past examples where we know both the 00:44.520 --> 00:47.200 inputs and correct outputs. 00:47.240 --> 00:49.600 Okay, let me scroll down here. 00:49.600 --> 00:54.360 We have NP dot random dot seed 42. 00:54.520 --> 00:54.920 Here. 00:54.960 --> 01:06.720 The first step we're going to set the random seed for reproducibility means every time you run the code 01:06.720 --> 01:09.730 you'll get the same random numbers. 01:09.770 --> 01:16.370 Okay, so here we are getting the same random numbers every time we run the cell. 01:16.410 --> 01:26.570 Then we have size equals to two times NP dot random dot random from 101 plus one. 01:26.610 --> 01:33.930 This is used for generating 100 random numbers 100 rows and one column. 01:33.970 --> 01:36.410 So let me explain everything. 01:36.450 --> 01:39.210 100 rows and one column. 01:39.250 --> 01:39.650 Okay. 01:39.690 --> 01:40.810 It's very simple. 01:41.010 --> 01:51.770 The third line is price equals to three times size plus two plus NP random times 0.5. 01:51.930 --> 01:56.890 This is a linear relationship three times size plus two. 01:56.930 --> 02:02.290 This is a linear linear relationship between size and price. 02:02.290 --> 02:06.210 So the price depends on size to be productive. 02:06.210 --> 02:14.060 For example, if the size is equals to two three times two six plus two, the price would be eight. 02:14.060 --> 02:21.420 So as the size increased, the price increase multiplied by 0.5 for the random numbers. 02:21.460 --> 02:28.740 This, the NP random number generates random noise from a normal distribution with mean equal to zero, 02:28.740 --> 02:32.780 and standard deviation equals to one times 0.5. 02:32.820 --> 02:34.420 Reduces noise level. 02:34.620 --> 02:40.220 Adding noise makes the data looks more realistic, not perfectly linear. 02:40.580 --> 02:44.620 Okay, this is a simple mathematical things. 02:44.620 --> 02:50.340 In order to get and make our data looks more realistic. 02:50.500 --> 02:59.020 In short, this code simulates a data set where price depends on size in a roughly linear way with some 02:59.020 --> 03:01.060 random noise added. 03:01.100 --> 03:08.380 Okay, now let me explain the data visualization using and the data frame. 03:08.380 --> 03:14.660 Here we have PD which is Panda Pandas library dot data frame. 03:14.860 --> 03:21.900 We talked about the data frame in the previous videos, and inside it we have MP concatenate and the 03:21.980 --> 03:25.380 columns the MP concatenate. 03:25.620 --> 03:29.780 The size shape 101 the shape. 03:29.820 --> 03:40.300 The price is a shape 100 to 1 which is there the range and NP concatenate joins the columns side by 03:40.300 --> 03:41.580 side column wise. 03:41.620 --> 03:44.700 Okay, and the result would be a shape. 03:44.740 --> 03:48.780 Each row has a size and price array. 03:48.940 --> 03:54.260 Then we have the PD dot data frame. 03:54.620 --> 03:55.700 We talked about it. 03:55.700 --> 04:00.660 This converts the numpy array into a pandas data frame. 04:00.700 --> 04:01.140 Let. 04:01.300 --> 04:03.780 Let me put this note here. 04:03.780 --> 04:05.540 So did you remember pandas? 04:05.740 --> 04:10.140 This is used for table like structure visualization. 04:10.140 --> 04:15.100 Also we have the two columns the size and the price. 04:15.100 --> 04:19.380 The size in square meters in thousands of square square feet. 04:19.620 --> 04:23.420 And the price would be in thousand dollars. 04:23.660 --> 04:28.470 Okay then we used to print the first ten results. 04:28.510 --> 04:31.110 Let me run and here we go. 04:31.430 --> 04:38.950 This is a random data generated using the random MP random. 04:39.150 --> 04:40.270 The numpy. 04:40.630 --> 04:45.990 We use the random number generation using the numpy. 04:46.310 --> 04:47.550 Also we have. 04:47.590 --> 04:54.750 We have the size and the price size is depending on random numbers plus one shifting. 04:54.990 --> 05:02.710 And the price depends on size times three plus two plus this random number times 0.5. 05:02.750 --> 05:03.110 Okay. 05:03.150 --> 05:10.630 So this is the linear relationship between the columns and the rows and x and y axes. 05:10.670 --> 05:12.990 This is the size in square feet. 05:12.990 --> 05:15.030 For example the first house is 05:15.230 --> 05:21.310 1749ft². 05:21.350 --> 05:27.790 The price is $7,290.