WEBVTT 00:01.200 --> 00:06.900 Now, let us discuss about the third project that you need to create, which is the unsupervised cloning 00:06.900 --> 00:07.340 project. 00:08.380 --> 00:13.840 In this particular project, you will be finding out different clusters of data. 00:14.830 --> 00:19.580 So the data which we have is that the market segmentation data. 00:19.990 --> 00:25.210 So in this particular case, you can use any kind of clustering algorithms. 00:25.210 --> 00:29.890 You can use game means you can use the rescan or you can use hierarchical clustering. 00:30.890 --> 00:37.520 In this, the problem statement is that interacting with strangers on social networking estimates has 00:37.520 --> 00:44.690 become a rite of passage for teenagers around the world, and many millions of teenage customers use 00:44.690 --> 00:52.610 this site and have attracted attention for marketers struggling to find an edge in an increasingly competitive 00:52.610 --> 00:53.080 market. 00:53.270 --> 00:59.070 So they are trying to find out what is of interest for these students. 00:59.090 --> 01:05.240 I'm finding out segments of students in what in the market and what they would be like. 01:06.050 --> 01:12.200 So for this, we have this particular data set where we have the graduation year of the student, the 01:12.200 --> 01:19.430 gender of the student age, the number of friends they have on Encinas and their likings like boyfriends, 01:19.430 --> 01:23.980 basketball, football, soccer, so softball, volleyball and so on. 01:24.260 --> 01:27.530 So there are a lot of interests which these students have. 01:27.800 --> 01:35.780 And based on these, these people are trying to segment these students into different market customer 01:35.780 --> 01:36.300 segments. 01:36.890 --> 01:39.060 So this is what you have to implement. 01:39.380 --> 01:46.970 Now, the first thing which you can do is you can find out the number of clusters by implementing gaming's 01:46.970 --> 01:58.430 or clustering or DV Skåne, or you can apply Bazzani and visualize this data on two components using 01:58.430 --> 02:02.030 DME and find out how many clusters are formed. 02:03.070 --> 02:10.210 And based on the number of clusters that you get from this Disney, you can actually decide on the number 02:10.210 --> 02:16.620 of clusters or you can simply use the elbow method in case of game means I'm used to seeing. 02:17.350 --> 02:21.450 Now, the choice of number of clusters completely depends on you. 02:21.700 --> 02:29.430 So you need to find out the best clusters form and you will be using the solid score for that. 02:29.680 --> 02:36.790 So you can use good index and find out how good the clusters are formed and compare different models 02:36.790 --> 02:40.240 comparing different implementations and. 02:41.390 --> 02:48.560 Create the project, so I hope you would have learned a lot from this unsupervised learning sessions 02:48.740 --> 02:55.340 and you will be able to create this project very easily as the implementations for unsupervised learning 02:55.340 --> 02:57.800 are very simple and straightforward. 02:58.130 --> 02:59.720 So thank you. 02:59.760 --> 03:07.370 And you can use these datasets on all the data sets which have been provided in different ways and use 03:07.370 --> 03:14.530 them for either classification regression or unsupervised learning and practice a lot. 03:14.840 --> 03:22.580 And once you have been practicing using these datasets, there are a lot of data sets present online 03:22.580 --> 03:26.060 in GANDIL or any other data universities. 03:26.210 --> 03:29.420 So all of these data sets are present online. 03:29.450 --> 03:35.150 So I hope you will try those on practice a lot using these. 03:36.220 --> 03:45.100 So as an ending note, I would like to say that this entire course was curated for you so that you can 03:45.100 --> 03:49.540 learn machine learning from the very beginning to the very end. 03:50.200 --> 03:55.930 You don't have to get back again to go through these documents again and again. 03:56.170 --> 03:58.080 I want you to learn all of this. 03:58.090 --> 04:04.570 I'm known for Fullwood and learning about new topics instead of going back again. 04:04.580 --> 04:12.190 That is why I have explained everything here from the deep mathematics behind all the algorithms to 04:12.190 --> 04:13.150 the statistics. 04:13.310 --> 04:19.150 So I hope you have gained a lot from this particular obsession and this. 04:19.960 --> 04:20.710 Thank you. 04:20.960 --> 04:23.520 I have a great time learning. 04:23.890 --> 04:24.460 Thank you.