WEBVTT 00:01.310 --> 00:05.360 Next topic is part of speech tagging. 00:06.200 --> 00:08.240 So what is here, the tagging me? 00:09.200 --> 00:17.630 Let's go to this linguistic feature of this spacy and signing some kind of pact to individual bogon. 00:18.500 --> 00:24.570 And those tax here are like a known pronoun, adverb, adjective. 00:25.190 --> 00:32.300 So those type of English grammar, when you assign me to some token because English is like an Merriman's 00:32.300 --> 00:39.680 exception language and all those human languages are all natural languages are all exceptions exist. 00:40.010 --> 00:42.800 So the same token might behave like an adverb. 00:42.830 --> 00:46.380 Sometimes that might be Lycan noun sometimes. 00:46.610 --> 00:54.020 So that is the beauty here, that once you assign it to some particular tag, you will be able to read 00:54.120 --> 00:56.270 those particular token differently. 00:56.360 --> 00:59.120 When you go further, further analysis of your tax. 00:59.720 --> 01:05.510 So Spacey has already in Belka, those part of speech tagging got implemented. 01:06.670 --> 01:13.190 For illustration purposes, I kept text like people is looking at buying UK stock for one billion dollar. 01:13.870 --> 01:19.790 Let's load it and let's load our simple English small size model. 01:20.950 --> 01:26.700 So first, Todd Bridges tried to connect it in his lies and allocate some space to. 01:29.010 --> 01:36.510 Now, once we apply this NLB function, let's try those quartos, so let me remove this one. 01:37.810 --> 01:39.110 Let's say not. 01:40.320 --> 01:42.670 And I'm just going to supply Presswood. 01:44.450 --> 01:49.420 And let's see what every single token so Kocan in, let's say. 01:49.710 --> 01:51.950 OK, let's think. 01:55.380 --> 01:58.010 Can not, let's say, Texas. 01:58.970 --> 02:00.290 So we get idea. 02:00.530 --> 02:06.860 What other tokens and then we'll apply those post tagging on each and every token. 02:08.440 --> 02:19.030 Let's say Goken not cause underscored, and that is a one more like op ed underscore certain Tolkan 02:19.630 --> 02:20.780 not tag. 02:21.100 --> 02:27.180 And this could and if you want a detailed explanation about each and every dose tag. 02:27.460 --> 02:35.250 You can use like us facey not explain dog tag from the school. 02:35.740 --> 02:39.000 So it will give you a detailed description related to each individual. 02:39.010 --> 02:41.330 Pocan has been assigned to other Tolkan. 02:43.560 --> 02:51.570 And you can see how police were known and ENPI and its description will be known proper, singular, 02:52.530 --> 02:53.610 this ease's LAGOONA. 02:54.310 --> 02:58.490 He x auxillary well, I guess per person, singular presence. 02:58.520 --> 03:00.180 So it's very much detail. 03:00.270 --> 03:01.980 Englis Grammatical stuff. 03:02.550 --> 03:06.840 Nahlah Like a symbol and it's a kind of currency. 03:07.470 --> 03:11.940 One is no kind of cardinal number billion. 03:12.000 --> 03:13.680 Billion is also no cardinal. 03:13.680 --> 03:13.920 No. 03:13.980 --> 03:20.940 So that is how beautiful and intelligently space he has identified each and every token UK like on a 03:20.950 --> 03:23.670 proper noun, proper singular noun. 03:23.910 --> 03:25.860 Startup is also a startup is a noun. 03:26.940 --> 03:27.330 All right. 03:27.360 --> 03:31.110 So that is how you can get the horse tagging. 03:32.950 --> 03:40.090 Now, if you want to get how many of them are Liko now and how many of them are a proper noun, if you 03:40.090 --> 03:46.330 want to do such a kind of analysis, you can just simply use dot dot count by metter. 03:48.390 --> 03:50.790 And here you can pass spacy. 03:53.150 --> 03:53.440 Not. 03:55.240 --> 03:55.810 Bopp's. 03:56.880 --> 03:57.480 Mattos. 03:59.360 --> 04:00.870 What was? 04:02.710 --> 04:03.800 Antiquities display. 04:05.550 --> 04:10.080 We'll be able to see it as displayed like a eighty five plus two time eighty seven accords one time. 04:10.410 --> 04:11.100 But what is this? 04:11.160 --> 04:11.880 Eighty five. 04:12.000 --> 04:13.290 Seven ninety two. 04:13.590 --> 04:14.850 So let us a little what. 04:15.810 --> 04:21.250 So far he tanah well in. 04:22.320 --> 04:24.330 So I'm just going to agree or item. 04:26.820 --> 04:27.530 And let me. 04:27.730 --> 04:33.200 This place was he and red and then we'll explain the meaning of all of them. 04:35.260 --> 04:36.370 I guess it's items. 04:37.840 --> 04:38.860 Yes, it's an item. 04:39.890 --> 04:40.400 Suppose. 04:40.490 --> 04:41.890 What is the meaning of 96? 04:41.960 --> 04:44.330 And that occurs to time if you want to grab it. 04:44.960 --> 04:48.610 We can use like a dog dog walker. 04:50.850 --> 04:52.580 And yet I can pass like an key. 04:53.790 --> 04:55.470 And we'll get much more ideas. 04:56.080 --> 05:02.910 How to apply tax also, because then it will just display the full object name only. 05:04.450 --> 05:10.130 All right, so PR., of course, to name a U.S. subsidiary, because when buying vlachos two times. 05:10.490 --> 05:12.920 So it goes kind of analysis also very much helpful. 05:13.520 --> 05:15.370 How many entities are there? 05:15.380 --> 05:16.640 How many votes are there? 05:17.030 --> 05:18.380 How many names are there? 05:18.440 --> 05:20.210 How many symbols are there? 05:20.540 --> 05:26.540 So where our document is much more oriented, let's say if we are dealing with some financial document. 05:26.570 --> 05:30.800 Maybe you'll encounter more and more like a money related stuff. 05:30.830 --> 05:35.450 So more and more symbols like our dollars sign other countries currency. 05:35.480 --> 05:38.990 You'll be able to see there are more numbers you will encounter. 05:39.410 --> 05:39.680 All right. 05:39.800 --> 05:43.910 So that is how it will extract those kind of features. 05:46.290 --> 05:53.490 Now, suppose this part of speech tagging, you want to display in some Raziel manner, we can use it 05:53.490 --> 05:54.820 from Spenceley. 05:57.120 --> 06:00.460 There will be a one song shine like a display. 06:01.620 --> 06:02.670 And let Sironi. 06:04.280 --> 06:06.010 Now, we already have a dark object. 06:06.050 --> 06:08.330 So we are going to use this display, see? 06:09.270 --> 06:16.770 Not Wrangler, which will display all those part of speech training assigned to each and every individual 06:16.830 --> 06:18.950 token in some manner. 06:19.990 --> 06:22.150 So first argument requires locks. 06:22.500 --> 06:24.120 And then we require Stice. 06:24.400 --> 06:25.740 So docs will be over. 06:25.960 --> 06:26.670 No object. 06:26.760 --> 06:27.690 What we accreted. 06:28.530 --> 06:30.180 Let's make it style. 06:32.130 --> 06:32.890 Or style. 06:32.910 --> 06:33.820 Just let us make it. 06:34.280 --> 06:39.190 BP only let's make Jupiter is equal. 06:40.740 --> 06:41.630 And let's run it. 06:43.200 --> 06:45.900 And it's very big, actually. 06:46.410 --> 06:49.590 Let's make it a little smaller because our sentence is not that much big. 06:50.280 --> 06:54.160 So Apple is better known than Easy's auxiliary. 06:54.780 --> 07:02.370 Looking is like an evil mind is also well, and there is a connection that something is looking for 07:02.370 --> 07:02.880 something. 07:03.300 --> 07:06.320 So those kind of connection also it has festivals. 07:07.320 --> 07:09.030 Let's make it a little smaller. 07:09.060 --> 07:09.790 So for that. 07:11.020 --> 07:13.150 We can pass on options like. 07:15.860 --> 07:16.970 Let's say distance. 07:19.220 --> 07:21.170 Let's make this 10 some handshake. 07:23.360 --> 07:26.990 And let me executed a search, it's a little smaller. 07:27.590 --> 07:28.950 Let's make it even smaller. 07:29.120 --> 07:29.540 80. 07:31.320 --> 07:31.770 All right. 07:31.800 --> 07:37.860 So you can see very beautiful results, all those part of speech tagging to every single tokenism really 07:37.860 --> 07:39.660 well in the public document object. 07:40.440 --> 07:40.770 All right. 07:40.800 --> 07:43.680 So that is all about the part of speech tagging. 07:43.780 --> 07:45.070 See you in the next video.