MultiClass Classification Approaches ll One Vs All and One Vs One Classification Explained in Hindi

Myself Shridhar Mankar an Engineer l TH-camr l Educational Blogger l Educator l Podcaster.
My Aim- To Make Engineering Students Life EASY.
Instagram - www.instagram....
Playlists :
• 5 Minutes Engineering Podcast :
• 5 Minutes Engineering ...
• Aptitude :
• Aptitude
• Machine Learning :
• Machine Learning
• Computer Graphics :
• Computer Graphics
• C Language Tutorial for Beginners :
• C Language Tutorial fo...
• R Tutorial for Beginners :
• R Tutorial for Beginners
• Python Tutorial for Beginners :
• Python Tutorial For Be...
• Embedded and Real Time Operating Systems (ERTOS) :
• Embedded and Real Time...
• Shridhar Live Talks :
• Shridhar Live Talks
• Welcome to 5 Minutes Engineering :
• Welcome To 5 Minutes E...
• Human Computer Interaction (HCI) :
• Human Computer Interac...
• Computer Organization and Architecture :
• Computer Organization ...
• Deep Learning :
• Deep Learning
• Genetic Algorithm :
• Genetic Algorithm
• Cloud Computing :
• Cloud Computing
• Information and Cyber Security :
• Information And Cyber ...
• Soft Computing and Optimization Algorithms :
• Soft Computing And Opt...
• Compiler Design :
• Compiler Design
• Operating System :
• Operating System
• Hadoop :
• Hadoop
• CUDA :
• CUDA
• Discrete Mathematics :
• Discrete Mathematics
• Theory of Computation (TOC) :
• Theory Of Computation ...
• Data Analytics :
• Data Analytics
• Software Modeling and Design :
• Software Modeling And ...
• Internet Of Things (IOT) :
• Internet Of Things (IOT)
• Database Management Systems (DBMS) :
• Database Management Sy...
• Computer Network (CN) :
• Computer Networks (CN)
• Software Engineering and Project Management :
• Software Engineering A...
• Design and Analysis of Algorithm :
• Design And Analysis Of...
• Data Mining and Warehouse :
• Data Mining and Warehouse
• Mobile Communication :
• Mobile Communication
• High Performance Computing :
• High Performance Compu...
• Artificial Intelligence and Robotics :
• Artificial intelligenc...

[@maurvinshah6422]

There are two types in one vs one: symmetric and asymmetric
Symmetric: n(n-1)/2 i.e. AB, BC, AC which is 3 models
Asymmetric: n(n-1) i.e AB, BC, AC, BA, CB, CA which is 6 models

[@sahilchoudhary1477]

Thanks Alot man❤️❣️✨

[@adityadhote4213]

Thanks man, I find it hectic to write it in my notebook and you even mentioned it in comment XDD

[@soumojitghatak8804]

ok sir

[@milesplays7518]

6 is the total number of classes present in the training pair that means if we take red green and blue we have to take take (red,blue) in first pair then (red ,green) in the second class pair and (blue,green ) in the 3rd pair and according to the formulae n(n-1)/2 =3 is correct but 3 pairs . And thank you sir for this beautifull expaination.!!!❤️❤️

[@Flowereacer89]

Sir, This is Genius level thinking.
I was watching a lecture from Colorado Boulder University. I got nothing out of it. This is Pure Gold sir.

[@PrinceShah-dx5he]

truly said appreciate your review !

[@Manindersingh-vv1ft]

The only teacher i have ever seen in my life who teaches simpler topics in simpler way and also complex topics in simpler way. By the way Sir I have one question for one vs one for n(n-1) /2 it will genarate 3 classifiers as if n=3

[@sushilaggarwal154]

Seems like instead of 3, he wrongly calculated 6.

[@varunbehere2072]

Yeah I was also confused with this
@@sushilaggarwal154

[@mishriti]

Very nice explanation...bt for one vs one for n(n-1)/2 it will generate 3 classifiers as if n=3

[@shivanshsingh5555]

same thing i was posting...it will be three

[@suryakantkadam7405]

Right.

[@AbhishekKumar-rt8xp]

Yahi to hum soche ki ye kya bol raha

[@iramarshad700]

I was thinking the same. 3 classifiers need by a combination of 2.

[@pratikshinde7817]

take n!(n-1)/2 it will generate 6 classifiers

[@rohitprasad908]

Sir if written formula is right then no. of classifiers generated will be 3.
Thank you sir

[@aasupatel5383]

Sir....according to u for number of classifiers (NumClasses * (NumClasses - 1)) / 2..for one vs one ...it will generate 3... if n=3

[@sanjaysethiya3373]

Thanks for this video
I think there should be one correction in this video, i. e
In One v/s One there should be three classes instead 6 classes

[@utkarshkalbhor3564]

nice explanation understand better than another channels
🙏👌

[@DbCo0pEr]

how to get these first three emojis?

[@anneswabose2166]

very nice explanation ... u really cleared my doubt... thank you so much for uploading this video

[@kaushalendrarathour9909]

Sir I've watched your maximum possible videos. So I think no one is better than you.
And now we need the video of "find S algorithm" & "candidate elimination" video.
🙏If it's possible so please sir this is my humble request pls make this video🙏

[@adityabhide6300]

It's 3..??

[@41abhishek]

In case of one vs one.. each classifier will go for identifying 3 classes? class A /class B / none of them.... Otherwise, during training.... how the un-matching class will be labeled?
Or..... in case of applying one vs one in multi class classification, should we need to filter two class data from the training examples and apply it to the corresponding classifier?

[@sanjuktamohanty4931]

So nice explanation. Got it clearly. Thanks a lot

[@syedamirali9144]

Level ha bhai . Bhai 1 chez sumjh ni ai mujy wo ye ka one vs one ka jo formula ap ny ider bataya ha us may apny bataya ka jitni classes han like 3 classes the C1 C2 C3 tu n(n-1)/2 = 6 Models ate but 6 /2 ho kar 3 ni aye ga ider

[@tejajesus4616]

n(n+1)/2

[@santoshsahu-oy5vo]

Thanks again, for your time and video 😊

[@naveedanjum3713]

according to formula answer should be 3 not six. Or it will only be 6 if formula is n(n+1)/2? am i wright

[@pankajsonawane7490]

you are right

[@rushikeshjawalekar9711]

Yes you are left

[@nehlmemon4086]

much helpful video, actually helped for exams and i have liked n subscribed too

[@anushkaupadhyaya3238]

Firstly [n(n-1)/2] = 3 in the above case, now the question is this means on 3 classifiers will be generated, then how to break 3 into pairs of classifiers ?

[@stellaquickdust8623]

C1c2
C2 c3
C1 c3

[@HrushikeshPatilOfficial]

c1c2 , c2c3 , c1c3

[@cbkapil]

dude you cleared my doubt well .. i think i got my answer ..but actually a research paper was using 8 features with hmv knn and naive bais. classififcation .. but i m unable to understand how c means and k means will classify the the 8 features as they cant as the both are unsupervised algorithm ??answer plz

[@neerajdewangan1990]

In one vs one n(n-1)/2 leads to 3 again which is same as one vs all. So what is the difference in both methods
Or else it is n(n+1)/2 =that will lead to 6 so plz confirm whether it is - or + in between.
Video is helpful for sure,👍

[@nehlmemon4086]

yeah you are right!! There is a mistake and i also got confused. Actually one v/s one will give rise to the 3 classifiers but the difference is that the classes will be considered in pairs forming the classifiers, C1 C2 will make M12, C1 C3 eill make M13 and the last one C2 C3 will make M23. but i must say this video was very helpful... thumbs up (Y).

[@santoshsahu-oy5vo]

Very Nice Dear...
n(n-1)/2 = 6 ? ye samajh ni aaya Sir, kaise hua? please help here.

[@aditiairen97]

sir there's a confusion in this video at the end. no doubt the way you explain is very good but plz explain the difference between 2 again

[@ASh-hb1ub]

Very Good Explanation Sir...👍👍👍👏👏👏❣

[@ahmedyar9517]

Great video

[@bluemonster5441]

Sir Reinforcement Learning pe video dalo please!

[@winviki123]

Awesome, thank you very much

[@earlymorningcodes6100]

Sir the formulae is n(n-1)/2=3 or formulae will n(n+1)/2=6 pls tell me sir

[@shaz-z506]

This video is really good. Could you please make a video on estimation-maximization and markow chain Monte Carlo.

[@shreyadutta1416]

Nice explanation , sir can you post video for semi- supervised classification only in theory format

[@ankurshukla7834]

Sir, could you plz explain one versus one with complete solving , because it's not available any where on Google

[@hrithiksethi4257]

This should be given with an example

[@akshayjinde511]

Thank You , got it, just make a video for parametric and non parametric learning

[@samratsinghrathore730]

What actually is a class here? I mean in the playlist you didn't explain previously too.

[@user-lg5lg3tt9m]

One by one me 6 classifier kaise hue? According to formula

[@kishorbankar5743]

As for one vs one:
in your example n is 3
as per formulae number of classifier should be 3 ((3*(3-1)/2)
but you have explain 6 classifier, How could be 6 classisfier???

[@mayuraugad1445]

Mast keep it up you are the best

[@belalvai6653]

N(N-1)/2 the correct formula and the answer will be 3 for one vs one classifier amount

[@mayurichandra6422]

Pls explain the purpose of doing this classification, how will this help in ML ? Pls clarify.. thanks

[@hassanjaved906]

sir, I m a little confused can u plz clear my confusion as why you are taking the probability of M1 0.90 in simple in what perspective you taking or model assumes that the probability of M1 is 0.90
or you are supposing this?
if you are supposing this is it important to take the probability of every model?

[@Noborotno.09]

Sir, at One vs One
n = 3
then, {3*(3-1)}/2 = 3 not 6(in the video 9:50)
Can you please clear this confusion?

[@sanketprajakta]

it is N(N-1) for all vs all

[@rishavanand1888]

nice explanation

[@Songvbm]

It is a great tutorial but I have a question.
What is to do if two classifiers in testing phase give positive response with equal probabilities?

[@masoudmaleki8455]

this is my question too

[@niveditasingh3096]

It's like choosing between two dish with same Price.

[@k-concepts1574]

Please correct the no of classifiers and their relevant impact on OF. Regards

[@hackme4435]

Detailed video on One vs one

[@pemessh]

Very impressive explaination.... Keep it up.

[@Howtosolveprobem]

What if the probability of positive classes is also equal then how results will be predicted? again a tiebreaker is observed then how we will classify this case?

[@niveditasingh3096]

How do we know/calculate the response type(+/-) and probability??

[@hometvfirestick]

Thanks.

[@anjuchaurasia3576]

It would be good if all topics have examples.

[@akshaychavan5511]

I think the second method's name should be All-Versus-All!

[@5MinutesEngineering]

Universally it is accepted as one Vs One.

[@akshaychavan5511]

@@5MinutesEngineering ohh, thought it should be All-Versus-All cause I read it in reference book! Bdw video was quite helpful. Thanks a lot.

[@sanjaisrao484]

Thanks

[@vedant6460]

great!

[@Flowereacer89]

sir one vs one bhi pura bata dete na plz

[@Aakash8739]

Very well explain

[@poojaingale428]

Thank uuhh sir very well explained sir

[@YashGupta-sg9kg]

a feed back try keeping video small and avoiding irrelevant talks.... rest thank you for making videos on such content

[@pranavsangawar3857]

Thanks for this awesome videos.. Sir plz avoid repeating things number of times....

[@riyapawar773]

In need of semi supervised classification video PLzzz do upload

[@prasad4u]

Sir, can we represent the confusion matrix with posteriori probability values?

[@vishaljaiswal4566]

Sir u r best

[@prakashSharma-kc8ek]

Same T-shirt 👕🌝

[@mAssAdroit]

Hi, for One-Vs-One if we need to create classifier then by your given formula n(n-1)/2 for 3 classes it will give you 3(3-1)/2 => 3*2/2 => 3, So how you generated 6 classifier, its confusing.

[@5MinutesEngineering]

Yes its 3.

[@minakshi_119]

Thank you sir

[@shahadatshajjat8503]

Well explained brother.

[@railwaypyas9411]

i understood the process but i confused that whats classifier? please help me somene

[@arshanabbas97]

Great man.

[@abhishekakash265]

sir ..kisi dataset me kitne classes create honge wo kaise pata chalega

[@somnathruidas648]

Sir wo posiblity kaise aya ?

[@arijitkaran9315]

Sir, what is tupple basically?

[@chetanmundhe8619]

very good

[@rima5210]

thnk u sir

[@ankitshivhare4567]

Sir Aapke one to one mein no. Of classes Ka formula galat hai ya explaination galat hai

[@utkarshkalbhor3564]

❤

[@padhaidotcom9276]

i am confused with f1, f2, f3

[@vyomveerudh2787]

How this process is implemented

[@muditmohantyagi]

what about binary classification

[@nikhilnambiar7160]

Need only on classifier then

[@johnwicckk]

Kabhi kabhi lagta hai ki aaphi bhagwan ho!!

[@adityamittal6284]

n(n-1)/2=3 par 6 bola h kaise?

[@5MinutesEngineering]

Yes,its 3

[@PRIYASINGH-kq9fc]

For a second I just get confused that did I forgot how to do simple calculations 🤦‍♀️

[@ayushtyagi1389]

Grt

[@tarleenkaur8415]

😊

[@DushyantSingh-dt]

sir one vs one m thoda problem hai....

[@ansatahir6923]

baqi to thek ha Jo ap ny .90 or baqi li ha wo samjh nahe Aya clear kren wo

[@user-sp3gk2me8e]

Hi

[@morningstarbbm]

n(n-1)/2= 6🤣🤣🤣🤣🤣 this kind of teacher are responsible for confusion.👍

[@Kishankanojiya.]

❤❤

[@deepak161091]

For one vs one, we require 3 models.

[@VishalYadav-nq1tv]

That's what I thought are we correct?

[@kishanpandey4798]

It will be ३ not ६

[@mohitmittal2627]

its confusing

[@soumojitghatak8804]

arey 6 nhi sir 3

[@benjaminbutton9779]

why are you in a hurry ? try to explain slower

[@rhutujbhoir2356]

[@vedant6460]

thanks

[@prasad4u]

Thank you sir.