you saved my life! wish me luck. I'll have an exam in the next 2 days. Hopefully, I can utilize all that you taught in this video. Keep on great work maim!
i think you did it for one iteration only; but in next iteration maybe any point can change its cluster as the two means are changed. So basically we need to reiterate same procedure unless cluster mean value does not change for two consecutive iterations.
Thank you Anuradha ji. Finally I understood what is K, what is mean, what is centroid, what is euclidean distance. Please create more videos covering major ML algorithms.
Hi mam, Good and neat explation for k- means algorithm it was very useful for me I need a explanation of CLARA and CLARANS in partitioning algorithm for my exam ..
Why you are updating Mean of cluster after every assignment? Aren't we suppose to update the mean after completion of single-one iteration according to the original Algorithm?
Formula of calculating Euclidean distance is needs to update as it contains (x-a)+(x-b) but it should be (x-a)+(y-b) also check square of 17, it should be 283 instead of 283.
This example is a bit difficult, you can simply take 2 rows directly and group it as 1 & 2, then find the Eucledian distance from each row and the shortest distance is your new row's group from 1 & 2, and the mean will be the grouped rows sum/2. Now the real concept who might go through my comment, K- means clustering is finding the way to group the similar set of data (of any type actually), then why we need a mean here? 1. When you calculate the distance from one point to other point you simply take a-b ( and you know that a>b, however this may not be possible in graphs or 3 dimensional plot, so you take square of sum of the distances for 2 values x,y and then you take a root so that if in a-b a
Shouldn't we update the centroid when we find all the distances between every data point and the previous centroid? you updated it with the average of the first two points only, why?
Are you sure your values are right ? At 9 min 31 secs the new K1 should be (185+179+182)/3 and (72+68+72)/3 . Also at 10 mins 20 secs k1 = (185 + 179 + 182 + 188)/4 and k2 = (72+68+72+77)/4 final centroid should be 183.5000 72.2500 for K1 and 169.0000 58.0000 for K2. I ran it in Matlab and Matlab confirms these answer.
I think mean is not calculated every step based on yours and her explaination. First , we assign all the data points to their nearest cluster and then take average of all the points in a cluster as a whole
Well explained. Just a minor suggestion. Most people watch on mobile so would be good to use entire screen rather than static title on left. I liked vedio
Hi mam, i want answer to this question -- Assume you are given n points in a D-dimensional space and a integer k. Describe the k-means ++ algorithm for clustering the points into k cluster
The centroid coordinates are continuously changing. Initially, we took co-ordinates of points 1 and 2 as two centroids, so should we not re-check if points 1 and 2 still belong to the initial cluster to which they were assigned?
Wow 💪 Now we will have DWM vidoes. thanks madam. please keep them coming, your content is helping us and yes the BDA paper was lengthy, but your vidoes covered 30 marks or more altogether. Page ranks, sums, FM Algorithm.
what is the point to calculate distance from centroid 1 to centroid 1 and from centroid 2 to centroid 2? isn't it obvious distance in this case gonna be 0 ?
Is this a correct approach lol? Don't we update the cluster centroid post assignment of all points and then iterate it till the clusters don't change. I am new to this. Enlighten me peeps.
.Thanks for your help .Im appreciate for your time ..But maybe there is a loose in Euclidean Distance [(x,y),(a,b)]=root of(x-a)^2+......(y-b)^2 and not ...(x-b)^2.as my point of view!!
![Hierarchical Agglomerative Clustering [HAC - Complete Link]](http://i.ytimg.com/vi/Cy3ci0Vqs3Y/mqdefault.jpg)
![Hierarchical Agglomerative Clustering [HAC - Complete Link]](/img/tr.png)
![K-Mean Clustering [Single Dataset]](http://i.ytimg.com/vi/rESDfP2ik3k/mqdefault.jpg)
Finally somebody that actually show calculations every step! Thank you so much you have my like!
3 years and this is still very much useful! Thank you so much.
you saved my life! wish me luck. I'll have an exam in the next 2 days. Hopefully, I can utilize all that you taught in this video. Keep on great work maim!
All the best..
I Finally understand the math behind this process. Thank you for walking through with actual data. This helps tremendously!
Helped a lot. Can't thank this lady enough. Just a small correction : Distance is (x-a)^2 + (y-b)^2
Thanks.. That's the correction
Mam, the explanation was CRYSTAL CLEAR. Thanks! keep making these types of tutorials. It really really helps
Thanks for the informative video ! @3:15, the variable should be 'y', instead of 'x', (y-b)
okay.
You made this look so clear and understandable. I sincerely appreciate you for this all-important K-means computation video!
OMG, I've been looking for this for so long!!! You are the QUEEN!!!
thank you very much, I was really confused how the implementation of this algorithm would be but you made it really easy to understand.
are you mad..??
The best video seen so far on K-means
i think you did it for one iteration only; but in next iteration maybe any point can change its cluster as the two means are changed.
So basically we need to reiterate same procedure unless cluster mean value does not change for two consecutive iterations.
Prasad Nagarale agreed. I have one question.which centroid values we should consider for next iteration..
Thank you Anuradha ji. Finally I understood what is K, what is mean, what is centroid, what is euclidean distance. Please create more videos covering major ML algorithms.
Your way of Explanation is easy to grasp Maam, Thank you 😇
Hi mam,
Good and neat explation for k- means algorithm it was very useful for me
I need a explanation of CLARA and CLARANS in partitioning algorithm for my exam ..
Why you are updating Mean of cluster after every assignment? Aren't we suppose to update the mean after completion of single-one iteration according to the original Algorithm?
Really straightforward and easy to understand.
Formula of calculating Euclidean distance is needs to update as it contains (x-a)+(x-b) but it should be (x-a)+(y-b) also check square of 17, it should be 283 instead of 283.
This example is a bit difficult, you can simply take 2 rows directly and group it as 1 & 2, then find the Eucledian distance from each row and the shortest distance is your new row's group from 1 & 2, and the mean will be the grouped rows sum/2.
Now the real concept who might go through my comment, K- means clustering is finding the way to group the similar set of data (of any type actually), then why we need a mean here?
1. When you calculate the distance from one point to other point you simply take a-b ( and you know that a>b, however this may not be possible in graphs or 3 dimensional plot, so you take square of sum of the distances for 2 values x,y and then you take a root so that if in a-b a
Shouldn't we update the centroid when we find all the distances between every data point and the previous centroid? you updated it with the average of the first two points only, why?
www.datacamp.com/community/tutorials/k-means-clustering-python
exactly
Are you sure your values are right ? At 9 min 31 secs the new K1 should be (185+179+182)/3 and (72+68+72)/3 . Also at 10 mins 20 secs k1 = (185 + 179 + 182 + 188)/4 and k2 = (72+68+72+77)/4 final centroid should be 183.5000 72.2500 for K1 and 169.0000 58.0000 for K2. I ran it in Matlab and Matlab confirms these answer.
woaaaa really really help me to understand more about K-mean
Thanks a lot. Systematic explanation and crystal clear.
Really simple and clear to understand, congratulations!
Anuradha great work. No where I got this detailed explanation.Please try to do videos for deep learning algorithms in a detailed way like this.
Your vedio Easily Understand.... Very Nice ma'am
Thanks for being my Savior for 10 marks.
mean is calculated in a wrong manner ,we have to take avg of all value in our set when ever some new value is added..
you are correct
I think mean is not calculated every step based on yours and her explaination. First , we assign all the data points to their nearest cluster and then take average of all the points in a cluster as a whole
what is the point of finding the distance between the two initial clusters?
the pts themselves are the centroids for their respective cluster, right?
Ma'am, how do we know that we have assigned the initial 2 items in the right clusters?
Great explanation
Excellent Video....Thanka lot mam...you saved my time
Thank you, Anuradha for such a comprehensive example.
Anurag Yadav thank you so much
Thanks for nice explanation, it helps.
What was the relevance of (0,21.93) values. There was no point of calculating that.
Thanks for the amazing job.
Dear Anuradha, thank you so so much.
Thank you so much ma'am for amazing explanation!
why u didn't do update means as you did in single dataset video?
Thanks mam 👍🏼Jazak Allah mam
Best one from all others👏
Great explanation for K-means!
Thanks.
a big thanks to you for this wonderful explanation
Well explained. Just a minor suggestion. Most people watch on mobile so would be good to use entire screen rather than static title on left. I liked vedio
Ashutosha Nadkarni thanks
Hi mam,
i want answer to this question -- Assume you are given n points in a D-dimensional space and a integer k. Describe the k-means ++ algorithm for clustering the points into k cluster
The centroid coordinates are continuously changing. Initially, we took co-ordinates of points 1 and 2 as two centroids, so should we not re-check if points 1 and 2 still belong to the initial cluster to which they were assigned?
Very very useful 👍 Thank u so much......💫
In the next dataset means 3rd one the value should be 17 square should be 289 it is 283 Ik the answer is correct just for informing
cluster assignment for the first 2 clusters is an assumption though we can justify it by euclidean distance calculation
very very useful video thank you so much madam.
Wow 💪 Now we will have DWM vidoes. thanks madam. please keep them coming, your content is helping us and yes the BDA paper was lengthy, but your vidoes covered 30 marks or more altogether. Page ranks, sums, FM Algorithm.
Thanks,
Will surely put .
Uploaded, Hierarchical Agglomerative Clustering, and Apriori Algorithm.
Anuradha Bhatia yes madam, I got the notification :) thanks.
More following.
Anuradha Bhatia That's great. Waiting eagerly :)
Don't update cluster centroid after every assignment update it after a whole iteration (all assignments in one iteration is complete)
Can we taken randomly any two initial centroids ?
very clear explanation!
what is the point to calculate distance from centroid 1 to centroid 1 and from centroid 2 to centroid 2? isn't it obvious distance in this case gonna be 0 ?
Really Excellent Mam...
Ma'am in which cluster should we assign the coordinate(or data point) if the euclidean distance is the same from both the clusters?
You can assign it in any of the 2 in that case
Wonderful lecture Mam... For more videos, the link is not working... Kindly post more videos on Datamining. Thank you.
Great stuff, thanks for explaining.
very good session...but there should be X-a and Y-b.
Perfectly clear, thank you very much.
excelente canal muchas gracias por tu conocimiento , like y suscripcion.
15 square is 225. Not 255. Please focus on small mistakes. Mean calculation is also wrong.
Ich schick dich gleich ins Vakuum!
Great! Thank you for the video!
wouldn't the equation be = root( (x-a)^2 + (y-b)^2 )
(Maybe small mistake)
Thanks a lot, ma'am. Helped me for my exam!
ar 3:50 its y-b ( euclidean distance) and not x-b
Why didn't we recalculate the cluster center in K-means clustering for a single dataset
4:53 - 15^2 is 225, not 255
How can we find the K value for a large data set?
Wonderful lecture mam.. thank you
Thanks
thank you for good and detailed explained
Thanks for the explanation.. it's is very clear...
Again the formula needs correction, I don't know why you are not correcting while going through it.
Thank you mam for detailed explanation
Correct typo, 17 sqr = 289 (not 283)
great and easily understandable explanation.
M Anbazhagan thanks
Thank you so much mam,, love 😍 ❤️
Is this a correct approach lol? Don't we update the cluster centroid post assignment of all points and then iterate it till the clusters don't change. I am new to this. Enlighten me peeps.
It helped 👌..thank you..but a simple mistake in the video which i noticed is square(15)=225 dear ..255 written there...
I too realized, sorry and thank you...
Best explanation ever about k-means!
Thanks!!
4:23 wrong formula, you forgot the y and replaced it by x
Superb explanation. Thank u
.Thanks for your help .Im appreciate for your time ..But maybe there is a loose in Euclidean Distance [(x,y),(a,b)]=root of(x-a)^2+......(y-b)^2 and not ...(x-b)^2.as my point of view!!
Great explanation madam
thank to explanation for K-means and can you share any pdf file realted to K-means.
Fully explained
Thank you ma'am for a great lesson !!
Thank you so much ma. Very helpful video.
Thank you😊
apne galat centroid nikala hai
Love's urs lecture
Good one mam, Thank you
@Anuradha Bhatia can you help me to apply K-Mean Clustering in Localization using VLC
thanks for everything this video is very instructive...
İSMAİL KAR NANOKUNG Thank You
simply wonderful/easy to understand
Great ma'am.
What is cross minimization in clustering
Mem Cosine Similarity Ka Sum Ka Examples How to find.... Ek Vedio Bnaake Rakho.... Plz
Thank you 👏👏👏🙏👼