how to use DBSCAN in case of multiple features? Is there any technique to use only few features or all feature but less important with very small weightage?
How to select the best algorithm for the data by looking at the data? This the question that I faced in many interviews. Can you please make a video on it?
Hi sir, a great thanks from me. A general question sir, I have performed DBSCAN, Fuzzy, and K-means clustering, how would I suggest which algorithm is best for the data? If the dataset is quite mess, large scale 10k rows, and skewed with big amount of outliers
If we give Epsilon=1 then it will randomly draw a circle on a particular data point and make its a circle with radius 1 ,so the core point is also chosen randomly ??????
Hi Aman, Thanks for the explanation, but my doubt is how cluster can be decide which point needs to take as a core point? What is the math behind that?
For each point xi, compute the distance between xi and the other points. Finds all neighbor points within distance eps of the starting point (xi). Each point, with a neighbor count greater than or equal to MinPts, is marked as core point or visited.(copied from web as It was quicker)
Hello sir, Which algorithm works well for customer segmentation wrt Recency, Frequency, Monetory? And is necessary to apply all the algorithms that is Kmeans, Dbscan, hier to the dataset and then come yo conclusion.
noise points are not consider in any clsuters right??? if new data is added ,then that data points form a cluster around noise point and then that noise point is also includes in a cluster or not???.then accuary of algortm changes or remains constant???
I'm doing Business analytics course and I refer to you video for understanding. Plz keep up the great work of enlightening us.
Thanks Sumit. Good luck with your course.
Excellent Explanation!! Please upload more videos of this similar kind sir..
Thanks Valli. Sure :)
Again nailed the topic. This is amazing how simply you have managed to explain the the concept
Thanks Again Sumit. Please share with your friends who might get benefitted :)
Nice and sweet explanation. I shared with my friends. Thank you Aman
Thanks Thayyib
your explanation is amazing man... keep going!
Thanks a lot.
HI
its very nice the way your explaing the topics really i love it thanks for the video
You are most welcome. Pls share with friends as well
Thank you so much. This is clear and on point. Subscribed!
Thanks Luam 😊
Thank you for the detailed explanation!
Welcome
Underrated Channel, Plus one sub
Thanks a lot Bala. Your words are my motivation
Nice and brilliant class sir.
Great explanation. Thank you!
Welcome
Really very nice teaching.....
Keep watching Sarthak
Well explained
Thanks Faiza.
very good
best explanation
Excellent explanation 👌
Glad you liked it Raja.
how to use DBSCAN in case of multiple features? Is there any technique to use only few features or all feature but less important with very small weightage?
FINISHED WATCHING
Excellent explanation, but one question..how can we evaluate DBSCAN , is there any test like we evaluate k- means ckuster by silhouette test?
Yes Naresh, I ll cover it in my upcoming video.
lucid explanation
Thanks Shubham.
Thank you sir
Welcome Asres.
Please put something for deep learning like cnns rnns and examples for those
Thanks a lot..
Welcome.
How to select the best algorithm for the data by looking at the data?
This the question that I faced in many interviews.
Can you please make a video on it?
This can not be done upfront without digging deep into data however it also depends on many factors. I will explain in one video separately.
Sir please upload a video on PCA next. 🙏
I will upload Ranajay.
Hi sir, a great thanks from me. A general question sir, I have performed DBSCAN, Fuzzy, and K-means clustering, how would I suggest which algorithm is best for the data? If the dataset is quite mess, large scale 10k rows, and skewed with big amount of outliers
If we give Epsilon=1 then it will randomly draw a circle on a particular data point and make its a circle with radius 1 ,so the core point is also chosen randomly ??????
Hi Aman,
Thanks for the explanation, but my doubt is how cluster can be decide which point needs to take as a core point? What is the math behind that?
For each point xi, compute the distance between xi and the other points. Finds all neighbor points within distance eps of the starting point (xi). Each point, with a neighbor count greater than or equal to MinPts, is marked as core point or visited.(copied from web as It was quicker)
Can you do a playlist on computer vision feature extraction techniques like hog sift (svm+hog), etc
Hi Augustine, I will try to add. Thanks for suggesting.
Can you pls make video on birch algorithm? Plz sir
Thanks for suggesting Suraj gupta :)
Hello sir,
Which algorithm works well for customer segmentation wrt Recency, Frequency, Monetory?
And is necessary to apply all the algorithms that is Kmeans, Dbscan, hier to the dataset and then come yo conclusion.
Hi Anshu, RFM is a good basic point to start with however we should try to fit data with advance techniques.
noise points are not consider in any clsuters right??? if new data is added ,then that data points form a cluster around noise point and then that noise point is also includes in a cluster or not???.then accuary of algortm changes or remains constant???
Hi Ravan, Noise will not be part of any cluster in any case. There is nothing like "Accuracy" in unsupervised ML.
@@UnfoldDataScience thanks ❤️
Sir please upload a video on Spectral Clustering next.
Sir, I want to add another point, it will be really beneficial if you make a separate video on unnormalized and normalized spectral clustering.
Sure Nayan, thanks again.
sir doubt on stats why are we converting the skewed distrubution to Gaussian distrubution?
Hi karthick, this we do typically in regression models as that is one of the assumption.
Can you also explain Isolation FOrest
Will do Saurav. Thanks for suggesting.
I have a question, which algorithm to use in varying density if not DBSCAN?
Try with k-means or hier
GRANDEEE
sir if interviewer asks differnetiate blw centroid and core point.........how can we proceed?
In DBSCAN its all about, core/border/noise points. Centriod is defined in K-means not DBscan
what is eps again can spell out didnt really catch the pronocuation?
Hi David, can you tell me which part of the video.