With each sentence you formulate, you explain exactly what is important and each sentence deepens the statistical understanding of the listener. I also find it wonderful that you go into every line of the R code and explain it. Thanks a lot for this.
Thank you for all the precious lectures. I majored experimental psychology 20 years ago. Now as a business man, I start using statistics again with R. Your lectures are really helpful to restore what I learned before. If I met a professor like you when I was young, I may have been a data scientist. Hope your health and happiness. A business man from South Korea.
After Spending 4 months on self paced learning this one hour lecture has provided more clarity to me than all the time i had spent on clustering earlier. a ton of Thanks to you sir for your support. I learn a lot from your videos. P.S- one small request, please also guide us how to do clustering if the Data if mixed up with both numerical & Categorical variables.
Thank you so much for the wonderful lecture. Actually I learned your all lecture in just 15 mins. It was effective and informative. One of your student asked about the purpose of this clustering. I think a good example is going to be its usage in internet search. Let’s say you are looking for an apartment to rent. Of course there are a lot of apartments to look for. But once you inter your favorite conditions, the system will come up with hundreds of apartments, that is where the cluster is used and the suggestions are made on that basis. Another example is be predicting which countries for example produce similar wine. Of course there will be details about the taste, color, smell and the percent of alcohol, but if you have 1000 different types of wine from lets say 80 countries, then you may not be able to say what countries produce the same type of wine or one country is imitating the other country. Here you have two used clustering to to find out.
Your session helped me a lot for finding exact clusters in my research hats off sir please include me for any new session you created I am following all your videos are excellent I want to learn many things from you its just superb
Thank you sir for making me more understood about clustering. I have one question, If I have data with different country each having yearly data from 1990 ,1991..to 2020, at that case how to used time series clustering?
Hi, Thank you for all the videos, it really helps to understand and implement. Could you please give something on all categorical variables without response variable and datewise data. I am looking for clustering on categorical variables. thanks again
Sir its very fruitfull lecture for me .wonderfull work done by you but i am little bit confused.......... i cant understand the R codes given below used for scree plot........why use (i in 2:20) in for loop instead of(i in 1:22) because in this data set we have 22 rows and in next step plot(1:20, plz explain # Scree Plot wss
In the code 2:20 means you want a scree plot for 2 to 20 clusters. It doesn't have 1, because only one cluster for entire data is not very useful. Similarly on the upper side even 20 clusters is too much.
You can refer to this discussion: stats.stackexchange.com/questions/173636/clustering-of-very-skewed-count-data-any-suggestions-to-go-about-transform-et
sir, thanks for the Cluster Analysis video. Can I request if you could please put a video on Circular visualization for non-genomic data similar to Circos or Circlize where you have many tracks corresponding to barplot, boxplots, heatmaps, as well as correlation in one single chord diagram? There are not many tutorials available for non-genomic data in this area. It would be very helpful if you could put one. Thanks a lot.
Sir, thank you. If possible can you please put a video on circular visualization of heatmaps along with chord diagrams (e.g., to show correlation) with multiple annotation tracks similar to Circos/Circlize
I have used my research data and followed all steps. Everything seems good. I really appreciate you! But for the last line, two dimensional clustering plotting. I found this error. Error in princomp.default(x, scores = TRUE, cor = ncol(x) > 2) : 'princomp' can only be used with more units than variables Within cluster sum of squares by cluster is as follows. [1] 0.000000 0.000000 6.534092 0.000000 (between_SS / total_SS = 86.4 %) Please help me, to get out of it. Thanks
Hello Professor, Thank you for your effort, please let me how can I came to know for the next Zoom session. you make the analysis in R-studio like a piece of cake thank you for these tutorials. Eagerly waiting for your reply.
With each sentence you formulate, you explain exactly what is important and each sentence deepens the statistical understanding of the listener. I also find it wonderful that you go into every line of the R code and explain it. Thanks a lot for this.
Wow, thank you!
Dr Rai, Thank you for teaching the way you do. The clarity in explaining concepts is unmatched. 🙏🏾🙏🏾
You are very welcome!
Your explanation skills are just fabulous ...how you make it so easy to understand 🙌
Thank you so much :)
Sir can you please make one video on Detection of heteroscedasticity .
Thank you for all the precious lectures. I majored experimental psychology 20 years ago. Now as a business man, I start using statistics again with R. Your lectures are really helpful to restore what I learned before. If I met a professor like you when I was young, I may have been a data scientist. Hope your health and happiness. A business man from South Korea.
It's nice to read your comments, thanks!
This is a fantastic lecture, Professor. Definitely my favorite of the class so far. Thank you.
You're very welcome! :)
It is awesome and clearly explained. The "R" example was also a good choice to show the praxis. Thanks a lot!
You're very welcome!
After Spending 4 months on self paced learning this one hour lecture has provided more clarity to me than all the time i had spent on clustering earlier. a ton of Thanks to you sir for your support. I learn a lot from your videos.
P.S- one small request, please also guide us how to do clustering if the Data if mixed up with both numerical & Categorical variables.
Thanks for your feedback and comments!
thankyou professor for such a detailed explanation. Can't stress enough how crystal clear my concepts have become after watching your videos :D
You are very welcome!
Wonderful class Professor. You made learning cluster analysis easy and to understand one important use of PCA as well. Thanks a lot sir.
Glad it was helpful!
Amazing Sir, with a detailed example. Easy of understanding the concepts, and how to apply!!!
Glad it was helpful!
Dr. Bharatendra Rai's lectures are so good, that I wish I was his direct student :)
Thanks for comments!
Thank you so much for the wonderful lecture. Actually I learned your all lecture in just 15 mins. It was effective and informative. One of your student asked about the purpose of this clustering. I think a good example is going to be its usage in internet search. Let’s say you are looking for an apartment to rent. Of course there are a lot of apartments to look for. But once you inter your favorite conditions, the system will come up with hundreds of apartments, that is where the cluster is used and the suggestions are made on that basis. Another example is be predicting which countries for example produce similar wine. Of course there will be details about the taste, color, smell and the percent of alcohol, but if you have 1000 different types of wine from lets say 80 countries, then you may not be able to say what countries produce the same type of wine or one country is imitating the other country. Here you have two used clustering to to find out.
Thanks for sharing these really interesting examples!
Your session helped me a lot for finding exact clusters in my research hats off sir please include me for any new session you created I am following all your videos are excellent I want to learn many things from you its just superb
Thanks for comments! If you are subscribed to the channel, you will automatically get notification about a new session.
That was a detailed explanation of inferences to be drawn.Thankyou
You are very welcome!
Thank you sir for making me more understood about clustering. I have one question, If I have data with different country each having yearly data from 1990 ,1991..to 2020, at that case how to used time series clustering?
For time series clustering, refer to this:
th-cam.com/video/QrTmwqK6edc/w-d-xo.html
Thanks for your interesting videos that are very useful and practical to learn
My pleasure!
thank you very much sir for explain every little point in code
You are most welcome!
all thumbs up!
Thanks again!
For the cluplot, is it possible to have a 3-D plot with the third principle component as the third dimension? Also amazing as usual, thank you!
Thank you so much Professor ✅🙏
You are very welcome!
sir how to get company names in last plot that is clusplot
how to find which variable is component 1 and 2
Hi, Thank you for all the videos, it really helps to understand and implement. Could you please give something on all categorical variables without response variable and datewise data. I am looking for clustering on categorical variables. thanks again
Great suggestion!
Sir its very fruitfull lecture for me .wonderfull work done by you but i am little bit confused.......... i cant understand the R codes given below used for scree plot........why use (i in 2:20) in for loop instead of(i in 1:22) because in this data set we have 22 rows and in next step plot(1:20, plz explain
# Scree Plot
wss
In the code 2:20 means you want a scree plot for 2 to 20 clusters. It doesn't have 1, because only one cluster for entire data is not very useful. Similarly on the upper side even 20 clusters is too much.
@@bkrai ok thanks got it now
You are welcome!
Sir, Does skewness in the data affects the clustering??
You can refer to this discussion:
stats.stackexchange.com/questions/173636/clustering-of-very-skewed-count-data-any-suggestions-to-go-about-transform-et
This is gold!
Thanks for comments!
In # Characterizing clusters does the actual value and normalizes values correlate?
Yes, actual value and normalized values always correlate.
Can you please tell me which R package are you using?
Package used is 'cluster'.
sir, thanks for the Cluster Analysis video. Can I request if you could please put a video on Circular visualization for non-genomic data similar to Circos or Circlize where you have many tracks corresponding to barplot, boxplots, heatmaps, as well as correlation in one single chord diagram? There are not many tutorials available for non-genomic data in this area. It would be very helpful if you could put one. Thanks a lot.
Thanks, I've added this to my list.
Sir, thank you. If possible can you please put a video on circular visualization of heatmaps along with chord diagrams (e.g., to show correlation) with multiple annotation tracks similar to Circos/Circlize
Thanks, I've added it to my list.
Sir I have pincode and booking over the pincode but when I do cluster analysis my result doesn't come as I want
pincode is not really a quantitative variable. Note that for cluster analysis you need quantitative variable.
@@bkrai sir can i get your mail id so that I can discuss my problem with you
seemabharat@gmail.com
sir how to hang it with company names, please
Include this:
labels = mydata$Company
@@bkrai thank you sir got it
I have used my research data and followed all steps. Everything seems good. I really appreciate you!
But for the last line, two dimensional clustering plotting. I found this error.
Error in princomp.default(x, scores = TRUE, cor = ncol(x) > 2) :
'princomp' can only be used with more units than variables
Within cluster sum of squares by cluster is as follows.
[1] 0.000000 0.000000 6.534092 0.000000
(between_SS / total_SS = 86.4 %)
Please help me, to get out of it. Thanks
Hello Professor, Thank you for your effort, please let me how can I came to know for the next Zoom session. you make the analysis in R-studio like a piece of cake thank you for these tutorials. Eagerly waiting for your reply.
I'll upload it on TH-cam. I also plan to do a live session for people getting started with R. I'll share info with all TH-cam subscribers.
@@bkrai sir, looking forward to the videos soon on getting started with R
.
Thanks, meanwhile you can also use this:
th-cam.com/play/PL34t5iLfZddv8tJkZboegN6tmyh2-zr_T.html
Great sir
Thanks!
Good video sir
Thanks!
Sir, can I have your email address. I am having doubts in linear regression model
seemabharat@gmail.com
Thanks
You are welcome!
Nice
Thanks!