WEBVTT 00:01.340 --> 00:02.540 Hey, everyone, welcome back. 00:02.990 --> 00:09.890 So the new project on which we are going to work up on that Ashton review classification and mainly 00:09.920 --> 00:11.330 throughout this whole project. 00:11.420 --> 00:13.910 We are going to use this and indicate lively. 00:14.540 --> 00:14.870 All right. 00:14.900 --> 00:21.060 So let me first input the data set and we see our business problem that what we are going to achieve. 00:21.710 --> 00:23.330 And what is the classification? 00:23.420 --> 00:25.970 We are going to do so from the file. 00:26.420 --> 00:27.820 Just click on upload. 00:29.460 --> 00:30.570 And let me go to the. 00:30.860 --> 00:31.900 And before that stop. 00:32.290 --> 00:32.570 Right. 00:32.810 --> 00:37.010 I kept this restaurant and the could leave you start ESV Fine. 00:38.260 --> 00:38.590 OK. 00:40.370 --> 00:42.990 So you can see fine uploaded. 00:44.000 --> 00:48.840 And I hope you did this whole project to four different videos. 00:48.920 --> 00:53.910 So in this video, we'll be focusing on our business problem and importing all those later. 00:54.800 --> 01:00.640 Next lecture is very much important because where we are going to do all those cleaning step, whatever 01:00.660 --> 01:07.040 we learn in this NLB basic section, and then we are going to apply one very simple feature engineering 01:07.040 --> 01:10.100 model, which is nothing but a bag of or model. 01:10.550 --> 01:16.070 And then we'll be applying one classification algorithm, which is based on a base theorem, a nine 01:16.070 --> 01:17.210 Navys algorithm. 01:17.870 --> 01:18.260 All right. 01:18.350 --> 01:26.420 So let me put all those necessary and required libraries, just as these were like Nubby Mike, Rutley 01:26.560 --> 01:27.830 and Pangas Library. 01:28.340 --> 01:30.020 And let me just add undescended. 01:31.360 --> 01:31.730 All right. 01:32.210 --> 01:34.450 So let's just import our restaurant review. 01:35.680 --> 01:36.920 So we got. 01:39.120 --> 01:47.760 And this could CSC and let just import a restaurant reviews, not VSP, fine. 01:48.160 --> 01:52.850 So here the separate that will be again, Celesta. 01:54.550 --> 01:56.340 And let me just run it. 01:58.140 --> 01:58.580 All right. 01:58.630 --> 02:03.580 So we are able to successfully lead this fight, but one more argument. 02:03.610 --> 02:04.420 I want to push it. 02:05.160 --> 02:05.590 Be. 02:08.110 --> 02:10.750 Voting so that we just keep it. 02:11.500 --> 02:17.620 So why that is because sometimes you may observe that, you know, place of such a kind of textual data, 02:17.740 --> 02:19.950 you may encounter some level quotes. 02:20.200 --> 02:26.830 So those double code, it needs to be ignored while reading this CSP or kind of PSV. 02:26.900 --> 02:27.150 Fine. 02:27.880 --> 02:32.470 Let me put all the data in to the data frame object. 02:35.250 --> 02:37.170 Let's observe first few records. 02:40.610 --> 02:47.360 And you will see we have every loser available and each and every reviews has been assigned to label 02:47.360 --> 02:48.680 either zero or one. 02:49.650 --> 02:54.370 So if you win here indicates that it is a positive review. 02:54.390 --> 02:55.640 It is like by to use it. 02:57.510 --> 02:59.710 And Yellow indicates he had the negative review. 03:00.250 --> 03:01.660 User does not like it. 03:02.500 --> 03:07.230 If you want to see data from the taling will get idea how many. 03:07.250 --> 03:08.710 Go to Lacasse when a relevant. 03:10.460 --> 03:16.210 All right, so you can see the total from zero to 999 Lachman's thousand recordset are. 03:17.890 --> 03:22.190 Let's try to find out how many of them started one and how many of them are zero. 03:22.990 --> 03:28.600 So later, like and we are going to play this. 03:28.730 --> 03:31.750 Will you find this good KONE's? 03:34.680 --> 03:41.160 So we hope Total find out the code side belongs to Class Zero and find it, the codes belongs to class 03:41.160 --> 03:41.350 one. 03:41.370 --> 03:44.050 That means it's a quiet balance, let's say. 03:44.670 --> 03:48.810 And what object to force is to clean this all decks. 03:49.110 --> 03:55.410 And in the next video, we will see how to clean all those texts and build the model, which eventually 03:55.410 --> 03:59.970 going to put any weather review like this one. 04:00.170 --> 04:06.700 Are not genetically the reviews positive or negative based on this review tax. 04:07.140 --> 04:07.590 All right. 04:07.650 --> 04:08.840 See you in the next video.