WEBVTT 00:01.020 --> 00:01.890 All right, everyone. 00:02.460 --> 00:09.700 So now time to test our newly created order, handcrafted created this tasting let. 00:10.380 --> 00:16.230 So for testing purposes on both of this model, random forest and a support vector machine, I have 00:16.270 --> 00:19.330 created a tree tasting sample like this one is. 00:19.410 --> 00:19.830 Hello. 00:19.950 --> 00:22.110 You are learning natural language processing. 00:22.780 --> 00:24.280 Then test two will be hope. 00:24.380 --> 00:26.610 You are doing good and learning new things. 00:27.120 --> 00:27.780 Various tests. 00:27.800 --> 00:29.120 We will be a congratulation. 00:29.130 --> 00:32.260 You won a lottery, offered one million dollar to Klimek. 00:32.330 --> 00:33.630 Call on this particular number. 00:33.990 --> 00:39.750 So as a human, just by looking at this testing dataset, it looks like that the first two are not a 00:39.750 --> 00:40.140 spam. 00:40.150 --> 00:41.160 They are a hard message. 00:41.190 --> 00:43.340 There is nothing wrong you are doing here. 00:43.350 --> 00:47.040 You are not encouraging the user to take some action. 00:47.850 --> 00:53.820 I mean, in terms of money or anything, whereas this three best user that you won some lottery. 00:53.820 --> 00:55.830 So that looks like a kind of spam. 00:56.310 --> 00:59.470 Now our time to test with this to model. 00:59.640 --> 01:00.510 So what do we do? 01:00.900 --> 01:03.240 Let's apply on a fast classifier model. 01:03.270 --> 01:05.580 That is nothing, whatever random forest model. 01:06.240 --> 01:08.610 So let's do predict. 01:10.300 --> 01:13.170 And let me play based on sameway. 01:13.710 --> 01:15.880 Let me for all feet. 01:17.160 --> 01:20.220 And I'm going to apply for test two and three. 01:20.880 --> 01:21.760 Let me execute it. 01:22.530 --> 01:23.910 I think it's this plate. 01:25.060 --> 01:25.380 Whoops. 01:25.710 --> 01:27.760 This is not defined. 01:28.530 --> 01:31.760 All right, now it's on display, the result of all three. 01:32.070 --> 01:36.690 So what we can do, we can just raptly their own print function. 01:42.780 --> 01:43.330 All right. 01:45.120 --> 01:50.590 All right, so first one is ham, ham and his farm, so perfectly, just like us as a human, it has 01:50.980 --> 01:54.130 correctly predicted the first two are not a spam. 01:54.310 --> 01:56.220 The last one is so random for us. 01:56.280 --> 01:58.590 Well, fine on all three example. 01:59.200 --> 02:02.020 Let's try to apply the same thing with the SVM. 02:02.860 --> 02:06.970 So now for SVM instead of classifier, I'm just going to replace it with Asmir. 02:10.300 --> 02:10.680 All right. 02:10.870 --> 02:12.930 And let just executing all. 02:13.460 --> 02:19.410 So far, less VMI, so it has correctly predicted the first two out of ham and at last one is spam, 02:19.450 --> 02:21.360 because it's a increasing use. 02:21.370 --> 02:23.060 It I mean, it looks like he and his family. 02:23.100 --> 02:29.550 So they are claiming to call on this particular number to claim for your one million dollar likely win. 02:30.190 --> 02:30.580 All right. 02:30.610 --> 02:36.100 So that is how you can make a prediction, because I know we've made a prediction out, actually, does 02:36.100 --> 02:36.250 it? 02:36.250 --> 02:36.800 Only now. 02:36.900 --> 02:38.440 This is my cell. 02:38.720 --> 02:41.800 What created this particular messages? 02:41.950 --> 02:43.350 And let me help. 02:44.110 --> 02:48.630 And it looks fine on a boat and forest line support vector machine also. 02:48.830 --> 02:51.460 So, like so see you in the next video. 02:51.490 --> 02:54.970 We will see some more stuff related to natural language processing.