Silhouette Score for clustering Explained | Silhouette (clustering)- Validating Clustering Models

[@shaneglean217]

The explanation is so concise and simple, even in spite of the fact that english isn't your 1st language. Amazing work my dude.

[@aaryalad7775]

I'm glad i found a teacher like you

[@ankitsapkota3713]

Never expected the conceptual delivery of this level without random steps out of the air in this accent.

[@Pannafreestyle]

Sir you are the best teacher for DS ! Keep up the great work

[@saijaswanth3036]

Thank you for clear explanation

[@isaacyekoniya8067]

This is simple and well explained, thank you Aman!

[@UnfoldDataScience]

Most welcome!

[@JuliaCh47]

Thank you so much, best explanation I've seen on the Internet!

[@UnfoldDataScience]

Thanks a lot for watching.

[@kushik.naveen]

This is so useful for understanding & revising. Amazing work! Thankyou!

[@victorcardenas9621]

Simple but powerful, thanks and greetings from Colombia!

[@MyerNore]

Clearer than my professor, for sure

[@awd_junkie]

Clear and right to the point! Thank you for your work! Much appreciated!

[@UnfoldDataScience]

Thanks for watching Chris. Stay safe. Tc

[@Felipe90820]

Amazing explanation! Congrats and thank you!

[@Anon11229]

Too good...Way better than online courses that take thousands and thousands of money keep up the hard work

[@UnfoldDataScience]

Thanks Zaid.

[@Krishna-pm8ty]

Excellent Amen. Perfect explanation👏👏👏

[@mdyounusahamed6668]

very concise and simple explanation.

[@nicolaevasiliu9430]

Amazing explanations! THANK YOU !!! You are an excellent teacher !!!

[@RezoanurRahman]

You explained it so well even a 5th grader would get it. Its so easily explained!.

[@RoopaNandu-k8b]

Aman, your explanation is really good. Very precise and in detail

[@uchennanwosu5327]

Excellent illustration and explanation.

[@ShaliniK-s5m]

Ur Explanation is simple and understandable,TQ sir

[@KHakimin]

This is gold! thanks for putting it very simple to understand

[@UnfoldDataScience]

Glad it was helpful Khairul.

[@paulholzapfel1856]

Perfect explanation. Thank you Aman

[@UnfoldDataScience]

Welcome Paul. Your feedback is precious to me.

[@izb1275]

Great explanation !

[@ravisavaliya4333]

at first it was very difficult for me to understand this, but you explained this in a very easy way....THANKYOU AMAN SIR FOR YOUR EASY AND INTUITIVE EXPLANATION!!!

[@UnfoldDataScience]

It's my pleasure

[@lhadz7290]

Thank you so much for Sir, I am so glad I came across your video. I look forward to learn more from you. Thank you again.

[@UnfoldDataScience]

So nice of you

[@sanjithaamarathunga9949]

Thank you! Simple and well explained!

[@UnfoldDataScience]

Glad it was helpful!

[@samiayasmin5782]

Thank you so much. Simple and clean explanation.

[@blackyogurt]

Great Explanation . thanks

[@nandinik7816]

Always to the point...and you dont waste any time actually...

[@UnfoldDataScience]

Thanks again.

[@the.quantum.paradox]

Very simple and easy to understand!

[@nandankakadiya1494]

Amazing explaination thank you sir

[@UnfoldDataScience]

Welcome Nandan.

[@amoghbharadwaj9252]

beautifully explained!!! Thank u so much

[@UnfoldDataScience]

Thanks Amogh. Happy Holi.

[@Suno-99]

beautiful explanation, keep it up

[@omniscienceisdead8837]

you are the best bro!

[@liqs_xd]

Great video, thanks for your time!

[@UnfoldDataScience]

My pleasure!

[@keinestudiere-chillzone633]

Straight to the point. Thank you

[@UnfoldDataScience]

Thanks a lot for watching

[@dbapple]

thank you, straight to the point!

[@muhammedthayyib9202]

Clearly explained!

[@vallimuthaiyah5098]

Thank you sir for very clear explanation with example.. 👍

[@UnfoldDataScience]

Welcome Valli :)

[@anshuraghav5620]

Great video

[@sangeethaagoudar2175]

Thank You so much for this video sir

[@puodzius]

Very good explanation, thanks

[@UnfoldDataScience]

Welcome.

[@sajjaddehghani3280]

Thanks for the very simple and easy to understand explanation

[@UnfoldDataScience]

Welcome

[@YouTribalChief]

What a gem of an explanation you give bhai! Truly you make things so simple and understandable! You deserve more!

[@UnfoldDataScience]

Thanks a lot. Pls share with friends as well. Have a nice weekend ahead.

[@venkataprashanth9831]

great and simple explanation, a big clap to ur efforts

[@UnfoldDataScience]

Thanks Venkat.

[@sourabhpramanik3957]

Its really a simple and good explanation for anybody to understand...thanks for making this simple explanation video.

[@UnfoldDataScience]

Welcome Sourabh

[@shaheerrehman3151]

Thank you. Much love to you from Pakistan

[@UnfoldDataScience]

Thanks Shaheer, pls share with friends if you like the content.

[@akashpb4044]

Awesome ...plain and simple 👍🏼

[@UnfoldDataScience]

Glad to hear that Akash 🙂, please share with others as well who could be benefited from such content.

[@parvsharma8767]

Thanks bro for explaining so perfectly

[@UnfoldDataScience]

Welcome Parv.

[@shaikhuzma786]

Tqsm sir so well explained ❤

[@diyap7343]

This is amazing!

[@UnfoldDataScience]

Thanks Diya.

[@dosendaring]

This is very good explanation, thank you Sir....

[@UnfoldDataScience]

Welcome Yudy.

[@PranavVashisth]

verry well explained !!!!

[@UnfoldDataScience]

Thanks Pranav.

[@MohdRashid-hi7cp]

Very nice explanation

[@UnfoldDataScience]

Thanks Rashid.pls share with friends also

[@kamal_douma]

very very good explnation

[@theone-jz2rf]

keep up the good work!!!

[@UnfoldDataScience]

Thank you.

[@mohe4ever514]

Very well explained!

[@UnfoldDataScience]

Glad it was helpful!

[@legolife3012]

Thanks a lot for this simple explenation

[@UnfoldDataScience]

You are welcome

[@beprepare8128]

Very well explained sir

[@UnfoldDataScience]

Thanks and welcome

[@optimalsolutionprogramming1299]

Thank you for your help. Really great explanation!

[@UnfoldDataScience]

Welcome

[@bangarrajumuppidu8354]

excellent

[@UnfoldDataScience]

Thanks a lot.

[@SurajReddy-hx2dx]

concept is well explained . But what do we do after this step. You have only calculated the silhouette score for one point. Do we need to calculate for all the points? What is the range of values present for silhouette score? When is it good or bad?

[@stevemungai3542]

Crystal clear

[@kavyasharma4738]

how can i thank you brother . you explained this topic so easily.

[@neajinnovations8587]

I got all the concept clear, but please when is it appropriate to use silhouette distance? it when you have mixed data? or mix data is best for elbow method?

[@sandipansarkar9211]

finished watching

[@anisadillap.7131]

oke i understand, thankyou sir ...

[@UnfoldDataScience]

Welcome

[@bipulnath4602]

Amazing explaination..plz make a video on Dunn index.

[@UnfoldDataScience]

Thanks Bipul.

[@jeanhwang18]

Thank you so much for your explaination. I have a small question: after I got point i ' s silhouette coefficient, I calculate other two points in cluster A, then I take the average of the three silhouette coefficients, let's say it's NO.1. What we can learn from No.1? If No.1 is close to +1, it's good, it means data points in cluster A are most similar to each other, right?

[@dr.zliiiiii]

nice video

[@gnaneshgn8341]

Nice video Sir thank you. please make an video of EDA sir ..!

[@UnfoldDataScience]

Thanks Ganesh.

[@sampathvinaykumarreddymajj790]

Topic is very well explained sir .. But should we learn DB Index, Dunn Index, Jaccard score ... or only this Silhouette score is enough

[@UnfoldDataScience]

Thanks Sampath.

[@vishwass5103]

Hi Aman thanks for this explanation. Please explain about assumptions of k means clustering

[@UnfoldDataScience]

Assumption at high level is, you data has clusters and centriods.

[@megalaramu]

Hi Aman, I do have a general question in clustering -I have heard of people do clustering in the dataset initially and then train separate model for each cluster. my question is lets take an example of telecom churn - there will be many services like internet, Phone, Dish etc. is there anyway we can cluster like internet users in one cluster, phone service in one cluster like that.

[@chrisx742]

Thank you.

[@ranajaydas8906]

Sir please make a video on PCA . It would be very helpful

[@UnfoldDataScience]

Sure Ranajay.

[@SACHINKUMAR-px8kq]

Thanks you so much sir

[@UnfoldDataScience]

Welcome Sachin.

[@Live_life_with_no_regrets]

is there any way to run it in SPSS? Is it correct if I use the average of the Euclidean distance of the cluster?

[@mmmommm237]

Great, ty!

[@mahipatil369]

hello sir.. great video i understand this topic very well,, thank you so much... just want to ask did you make video of implementation of silhoutte score in python?? if yes plz provide link..

[@UnfoldDataScience]

Thanks a lot. Its very simple to do in python - see this link - stackoverflow.com/questions/59919627/how-to-calculate-the-silhouette-score-for-each-cluster-separately-in-python

[@almonddonut1818]

Thank you!!

[@UnfoldDataScience]

Welcome.

[@veyselaytekin8734]

thank you

[@handsomemehdi3445]

Thank u for simpler explanation! A question: In our case, it seems that for both clusters, b should be the same unless we have more than two clusters, (or since those clusters are mutually nearest) doesn't it?

[@lkjhlkjh1954]

I think that’s why we do not calculate silhouettes score for less than 2 cluster

[@TheGentlemanGR]

So which score is better? -0.2 or 0.21 ?

[@arisurya9867]

Nice video sir. But i have problem when calculate cohesion in cluster just has one data point. How i should set value of cohesion? 0 or 1?

[@UnfoldDataScience]

Thats not a good clustering process, clusters need to be reformed.

[@arisurya9867]

Thank you sir 🙏

[@himanshumangoli6708]

As we have find cohesion and separation for only one point in one cluster A, so we have to find Silhouette score for all points for all points in same cluster so just to ensure that points are correctly clustered ?? Please explain??

[@UnfoldDataScience]

Yes , one point is just for example.

[@xendu-d9v]

thanks sir

[@UnfoldDataScience]

Most welcome

[@prashanths4455]

super bro

[@UnfoldDataScience]

Thanks Prashanth

[@subhasisdutta3711]

you have said "Min distance of A1,A2,A3 is the value of b ". Here you are talking about mean value of three or the which one has the minimum value among three , we'll chose it. Please clear it once, if possible

[@UnfoldDataScience]

HI Subhasis, A1, A2 etc are "mean" Distance from various other clusters ok. Now let's take att these A1 TO AN values and take the minimum from this set. This mimum is b.

[@subhasisdutta3711]

@@UnfoldDataScience thanq for the clarification. Have a blessed life ahead

[@christopherlegarda5164]

This guy is god!

[@parikshitgurjar5545]

hello sir, Plese make a video on the explanation and implementation of ppf,pdf,cdf in python with simple explanation and understanding, I am very confused in this.

[@UnfoldDataScience]

pdf cdf video is available Parikshit. Please search in my channel.

[@parikshitgurjar5545]

@@UnfoldDataScience Thankyou sir❤️

[@pruthvi491]

How to get silhouette score for spectral clustering

[@UnfoldDataScience]

I Will check.

[@gretchenbekker2371]

How do I get the amount of data points within the clusters ?

[@UnfoldDataScience]

Levels can be found out after clustering then we can count using pandas.

[@salilshukla08]

thanks

[@UnfoldDataScience]

Welcome.

[@iftikhar58]

clear

[@UnfoldDataScience]

Thank you

[@abeerhamid]

How can i read the graph of silhouette score? Most important thing you didn't tell

[@UnfoldDataScience]

Positive Scores (close to 1): Samples are well-clustered, and there is good separation between clusters. Scores around 0: Samples are on or near the boundary between clusters, indicating potential overlap. Negative Scores: Samples are likely in the wrong cluster.

[@nerdymath6]

is 0.4 is good silhouette score?

[@UnfoldDataScience]

Dont think so

[@dariomelconian9502]

Is 0.39 a bad score?

[@UnfoldDataScience]

can be better

[@sameeransari9129]

apki english samajhne me time lg jata hai.. ki aap kya bolna chahte ho

[@sameeransari9129]

Sir aap hindi me samjhaya karo toh zyada aacha samjha payoge...