WEBVTT 00:01.230 --> 00:08.180 In last session, we discussed about gaming's clustering and we had already implemented agglomerate 00:08.200 --> 00:10.280 clustering using the iris deposit. 00:10.650 --> 00:18.510 So let us continue further and use the same dataset to further implement gaming's clustering and see 00:18.510 --> 00:19.680 how it is different. 00:20.310 --> 00:28.530 So in this for this implementation, we will import gaming's from Escalon cluster after this. 00:28.650 --> 00:36.110 We have imported the latest dataset and we have already dropped the unwonted columns. 00:36.120 --> 00:39.480 That is the simple idea. 00:39.480 --> 00:40.710 And these species. 00:42.480 --> 00:51.620 After that, we have scaled the data and printed the Dendrobium and linkage for the same for the details, 00:51.630 --> 00:58.700 who can view the previous implementation of hierarchical clustering later on the clusters? 00:58.710 --> 01:04.260 And we found out that the vast number of clusters came to be. 01:10.550 --> 01:16.820 Forty three for three, so we used that particular clustering algorithm. 01:17.360 --> 01:24.410 Next, we will model the clustering model so we will find out the clusters for the end cluster equal 01:24.410 --> 01:24.980 to three. 01:25.310 --> 01:30.520 And then we printed the same and created a plot for the same. 01:30.890 --> 01:33.410 And we got this particular plot. 01:35.140 --> 01:41.990 Now for finding out the gaming's cluster, we have imported Kamins from Escalon factory. 01:42.700 --> 01:49.930 Now we are again testing for any clusters ranging from two to 20, and we are running a low on key, 01:49.930 --> 01:52.300 which is valued between two and 20. 01:52.570 --> 01:58.750 And we are running game is for each of these values, for running game means we only provide the number 01:58.750 --> 02:02.070 of clusters and the data in this data. 02:02.200 --> 02:07.800 I have provided the sample and then set Bellbird for actual implementation. 02:07.810 --> 02:16.270 You will use all the columns of a job prison because that will provide better clusters and it will be 02:16.270 --> 02:17.270 more helpful. 02:17.770 --> 02:23.830 While this is just for visualization purposes and wanted to make this a little simple for you born the 02:23.830 --> 02:24.220 son. 02:24.550 --> 02:27.130 So I have used only these two columns. 02:27.640 --> 02:35.470 So we will apply start Stopford and after playing games Rodford, I will bring the Senate scores for 02:35.470 --> 02:35.760 all. 02:36.220 --> 02:44.140 And here you can see the Senate scores game out to be these values out of which the best Sillett score 02:44.140 --> 02:47.170 is two, which is zero point four. 02:47.620 --> 02:55.090 But it also shows the second best to be the Senate score three, which is forty three point four seven. 02:55.810 --> 03:01.000 This is because we don't have we have not considered all the columns. 03:01.000 --> 03:08.140 We have considered only two columns I busway destroying two columns as not as two clusters as the best 03:08.140 --> 03:09.160 number of clusters. 03:09.550 --> 03:14.500 So let us create the clusters with the same higher level. 03:14.650 --> 03:21.640 One is the label which has been generated using the clustering and labelled is the one which is generated 03:21.640 --> 03:23.080 from Thicky means clustering. 03:23.590 --> 03:27.420 I have simply taken the values which Wychwood predicted. 03:27.430 --> 03:33.970 The labels, which were predicted using the key means clustering in the label too and the IT in the 03:33.970 --> 03:35.020 eye distinctness it. 03:36.870 --> 03:46.850 Here I have plotted the leak of the data, the plot for four separate versus a Zeppelin, I'm here, 03:46.860 --> 03:54.420 you can clearly see why two clusters are more better in comparison to three clusters in this particular 03:54.420 --> 04:03.060 scenario where we have considered only two columns and same thing is depicted here in case of the cluster 04:03.060 --> 04:04.470 created for agglomerate.