WEBVTT 00:00.590 --> 00:01.460 All right, everyone. 00:01.490 --> 00:07.250 So another machine learning algorithm, which we are going to play on a same spam message classification 00:07.280 --> 00:07.640 dataset. 00:08.240 --> 00:10.430 That is a support actor machine. 00:10.910 --> 00:11.900 So Ledger's a plate. 00:12.170 --> 00:19.640 And let's try to value it with the same classification criteria like accuracy score and confusing rhetoric. 00:19.640 --> 00:21.200 Center Classification Report. 00:21.260 --> 00:27.380 Let's try to generate so seem like earlier on live we are going to use it, but now instead of random 00:27.390 --> 00:29.780 fighters, it will be a support racked up machine. 00:32.290 --> 00:35.710 So first, let me quote from scalar. 00:39.170 --> 00:41.260 Not SVM. 00:43.910 --> 00:49.580 Let's import as we see and now what I'm trying to do. 00:49.850 --> 00:57.920 Let me select the full classifier object and just we are going to change it like a instead of random 00:57.920 --> 00:58.390 forests. 00:58.810 --> 01:00.890 Let us use s.E.C. 01:02.170 --> 01:04.530 And you just cannot supply this argument. 01:05.020 --> 01:08.070 So if you see the documentation of this as we see. 01:12.090 --> 01:14.730 It has created a full documentation. 01:15.390 --> 01:18.000 So with that, we are going to supply the core argument. 01:18.090 --> 01:21.110 Let's say the see arguing, let's just make it Hendrickx. 01:22.460 --> 01:28.370 And Gamma, so gamma will be let's just make it part of. 01:29.120 --> 01:32.830 And instead of classify it, I'm just going to use like a SVM object. 01:33.960 --> 01:34.960 And let me execute. 01:39.480 --> 01:39.900 All right. 01:39.930 --> 01:42.360 So there is some typo here he's equal to. 01:45.320 --> 01:50.730 I guess sees an unexpected argument that that is a capital is not a small city. 01:51.500 --> 01:52.570 Let me get any. 01:54.400 --> 01:54.820 All right. 01:55.270 --> 01:58.010 Now, let just trained is they don't know it pessimism. 01:58.850 --> 01:59.240 Martin. 02:00.010 --> 02:06.550 So for that, we are going to use this ship X and the school plane. 02:07.720 --> 02:10.340 And why this plane? 02:11.410 --> 02:12.680 And let me then you. 02:14.450 --> 02:17.750 All right, so our model got creative next days. 02:18.230 --> 02:19.800 We need to predict this thing. 02:19.910 --> 02:21.850 So I'm just going to select this one. 02:24.660 --> 02:28.140 And instead of classify it, let's put it like a SVM. 02:29.210 --> 02:31.940 So we got the result of a prediction. 02:32.690 --> 02:39.680 Let's try to find now accuracy score and we will be supplying the scope test. 02:40.390 --> 02:43.760 And let us see whether it got improved or not. 02:44.670 --> 02:46.070 It will be my prediction. 02:47.720 --> 02:48.170 All right. 02:48.200 --> 02:50.630 So we got almost similar accuracy. 02:50.660 --> 02:57.110 Ninety four point forty three in our earlier case, it was a ninety four point four. 02:57.120 --> 02:59.250 It's a little less than random forest. 03:00.020 --> 03:03.700 Accuracy, we got better somewhat similar accuracy we have achieved. 03:04.130 --> 03:06.860 Let's try to find a confusing matrix for that. 03:11.560 --> 03:11.980 All right. 03:12.100 --> 03:13.960 Almost the same leg earlier. 03:13.990 --> 03:16.120 We have a twenty five records. 03:16.220 --> 03:16.890 Got it wrong. 03:16.930 --> 03:20.290 But you know, it is it was 24 records. 03:22.170 --> 03:26.320 Whereas the remaining all records completely predicate. 03:26.670 --> 03:27.030 Right. 03:27.750 --> 03:31.790 You can even do seem like a classification report also. 03:33.800 --> 03:35.490 So let me pass here. 03:36.020 --> 03:39.120 Whiteface I and this quote, Brad. 03:40.380 --> 03:42.760 And it will give a detailed report. 03:43.460 --> 03:44.820 So we this mistake. 03:44.880 --> 03:47.550 Let's wrap it around the brain function. 03:49.330 --> 03:49.810 All right. 03:49.860 --> 03:54.320 So you can see that's calculated, this precision recall effort, schools, support for both of these 03:54.400 --> 03:55.150 getting high. 03:55.580 --> 03:55.760 Right. 03:56.010 --> 04:02.310 So nothing much improvement, almost similar accuracy we got with this SVM model also. 04:02.910 --> 04:04.280 But now we have two models. 04:04.340 --> 04:08.260 Let's say it will be a classified model that is nothing but other random forest model. 04:08.780 --> 04:11.270 And the other one is a has to be a model. 04:11.690 --> 04:18.530 What we'll do next next year, you will apply instead of this testing that I say will apply, our own 04:18.950 --> 04:26.840 created some spam and a harm message and we'll try to find whether it SVM our random forest. 04:26.960 --> 04:28.790 Will you correct this or not? 04:29.000 --> 04:30.090 So see you next video. 04:30.430 --> 04:32.470 I'll let handcrafted tasting gas. 04:32.750 --> 04:34.030 We will apply on top of.