I literally understood this thing just by watching this video ONCE. I hope you realized how good is your ability to teach. Who ever have chance to be your student is blessed.
I love the way you explain math using real world scenarios. Its makes us easy to understand and grasp the concept very easily. Thank you sir your are a good teacher even 5 years old kid will understand this algorithm very easily by the way you explained.
I have just finished watching it. Fully understood. Sir right from the word go I have watched all your videos and all of them are simple and easy to comprehend. Keep posting... Waiting for the next...
Kudos to the explanation. Your efforts are appreciated. But there are some correction to be made. 1) The Euclidean distance of Zaira is 29 and not 9 2) The Euclidean distance of Smith and Michael with respect to Angelina are equal 3) The 3 nearest Euclidean distance with respect to Angelina are of Names = (Sara , Smith , Michael) Euclidean Distances = (11 , 10.049 , 10.049) Sport Class = (Cricket , Cricket , Football) Anyways decision will remain the same (i.e - Prediction will be that Angelina will belong to Cricket Class) But what if the sport class were (Cricket , Neither , Football). As we have 3 category in the sport class so it is not a binary classification what will be the conclusion in the above mentioned case?
@@abhishek0001. Can be given in the question but if that is not given there is always another way to start. Count the number of possibilities there are in class attribute you are trying to predict (in this case it's class of sport). There are 3 possible options there { neither, football, cricket } therefore you start with 3. Do not make k less than the number of class values, meaning in this situation you would not make k=2. One thing that can happen--If 3 is not giving you a definitive answer for which sport Angelina is most likely to play, you must raise the value of k. Nearest neighbors are the records with the lowest distance. For example k=3, and the 1st nearest neighbors sport = neither, 2nd nearest neighbor = cricket, 3rd nearest neighbor = football, this doesn't tell you which sport Angelina is most likely to play as there is no repeating sport. This is when you raise the value of k. With k=4 and 3 possible class of sport options 1 sport must be more common than the others.
I really appretiate the indian utubers who guides us with full force...because of this. I usually try to search the indian guiders for any topic relate to study.
Ooh my God! I just understood this now because of you after so many trials to read.... May God bless you man! You Indians are a blessing to this Universe 🎉
Clearly explained with simple example.. within short span of time I cleared the concept.. but you have to include python programming for this algo also
Yeah ,So we Have to take values 10,10.05,11 which is from Arun-> Cricket,Micheal -> FootBall,Sara->Cricket. Where 2 are Cricket ,1 football .Majority is Cricket,Ans is Cricket.
Very clear. I'm not in this kind of things, but I think that it can be weighed in some ways. In this example it seems that the gender is not very important, because in the euclidean distance it doesn't really count compared to the age gap . If I want to give it more importance I can use male=0 and female=10 or male=0 and female=20.
Great video! Think Smith is 10.05 and from my understanding for distance based algorithms (such as kNN) it's best to standardized the data so that columns with larger ranges don't over impact results
Yogesh Ji, I have seen your vlog for the first time and really impressed with the ease of understanding you are inculcating in the explanation. Thanks and keep sharing such useful vlogs.
Thank for video with such clear, step by step explanation! There is one error: The distance between ZAIRA and ANGELINA is 29, not 9. So, the 3 closest distances to ANGELINA are: MICHAEL (distance: 10,05) - Football SMITH (distance: 10,05) - Cricket SARA (distance: 11) - Cricket
this is such a simple way of teaching like really with a pen and paper , I dont get it people using useless animations and out of the context examples making videos more complex to understand . This is the best way of retaining information to the memory also makes it way more crisp to understand .
Thanks for the video. That was good. What if the KNN prediction was not a majority one? Like in the above example if 3 candidates say 'cricket', 'football' and 'Neither'. Which one to choose ?
And what happens in the Ambiguous cases? Like in the resultant K values i.e three people with 9, 10, and 10.05 distance respectively... what if they have different likes like Cricket, Neither and Football respectively? There is no common choice. In case if we choose the `nearest` person who is here with 9 distance is the one, then what if there were two with same minimum distance and different classes ?
What if for 3 nearest neighbor we get three different classes, What we should choose in that case? Can you please clear this doubt... Python code will be appreciated...
How is Zaira's distance 9? I am getting 29. Clearly, even without calculating, we can see that Sara would be closer than Zaira. Can you please let me know if there is anything wrong with my understanding? Thank you in advance.
you are awsome only one doubt i have as you have take example for 2 features what if there are 3 or more then 3 then how we are going to calculate it. like what would be the formula
Brilliant. Thank you so much. Simple pen + paper made this so easy to understand. I have a question: say those 3 all had different classes. So Zaira had Cricket, Sara had Neither, Michael had Football. Is the solution here to increase the K number?
As the video is nice and easily understandable ,but it contains some mistake in calculation so kindly check it before uploading the video here bro/sir. And it too useful . Great representation👍👍
Thank you for the explanation. I have a small doubt. What if the final 3 nearest neighbors were from 3 different groups "cricket", "football", and "neither". In this case, there is no majority what should we do?
While applying knn ,it takes care to select the points avoiding such situations..coz data cleaning and preprocessing will be applied first..the example above is only explain the concept..if you want you can watch implementation of KNN
