Hello Suganya, thank you very much for this presentation of k-modes. I understand that it is not necessary to encode the data? Or maybe it is required in a particular case? if so, which encoding is the most pertinent (i.e. onehot, label, etc.). Thank you!
Thank you and K-Modes can process nominal data as such without encoding. If your dataset has combination of both discrete and continuous values, then K-Modes may not be suitable.
Hi Suganya -- Firstly, very well explained K-Mode Clustering. Now my question is how we will choose the optimal value of K in K-Mode Clustering like how we did in K-Means. I have tried using Elbow & Silhouette method in python but I can't calculate score value for K-Mode in python( for ex syntax --> kproto_1.fit(df_array).score(df_array)). Please advice which approach & how we will get to know optimal value of k in K-Mode Clustering in Python.Thanks.
Hello, Thank you for this presentation. Could you please give further insights on how to determine cluster numbers by using an algorithm in Python for k-modes clustering? It would be really beneficial if you kindly provide the code for the same.
Hi, Mam thanks for the video. I have a similar problem statement but the problem is I wanted to cluster the doctors based on hospitals to send personalized emails. There are over 10k distinct hospitals and there may be more than one doctor associated with each hospital. So here if I choose a cluster size of 10k for each hospital, is it a good idea as it might increase complexity?
Thank you Kab. You are a great teacher.
Glad you think so!
Good explaination mam ❤
Really Helpful. Thank you so much :)
You're welcome!
Thanks a lot mam for clear explanation ✨
Most welcome 😊
Hello Suganya, thank you very much for this presentation of k-modes. I understand that it is not necessary to encode the data? Or maybe it is required in a particular case? if so, which encoding is the most pertinent (i.e. onehot, label, etc.). Thank you!
Thank you and K-Modes can process nominal data as such without encoding.
If your dataset has combination of both discrete and continuous values, then K-Modes may not be suitable.
Encoding for Categorical values
ruclips.net/video/B1CU7s2Fktk/видео.html
Hi Suganya -- Firstly, very well explained K-Mode Clustering. Now my question is how we will choose the optimal value of K in K-Mode Clustering like how we did in K-Means. I have tried using Elbow & Silhouette method in python but I can't calculate score value for K-Mode in python( for ex syntax --> kproto_1.fit(df_array).score(df_array)). Please advice which approach & how we will get to know optimal value of k in K-Mode Clustering in Python.Thanks.
Clear explanation
Where do I get more information about k-modes?
GOOOGLE😶
Hello, Thank you for this presentation. Could you please give further insights on how to determine cluster numbers by using an algorithm in Python for k-modes clustering? It would be really beneficial if you kindly provide the code for the same.
Thank you!
Welcome!
Hi, Mam thanks for the video. I have a similar problem statement but the problem is I wanted to cluster the doctors based on hospitals to send personalized emails. There are over 10k distinct hospitals and there may be more than one doctor associated with each hospital. So here if I choose a cluster size of 10k for each hospital, is it a good idea as it might increase complexity?
it will increase complexity. The decision may be taken based on size of dataset
thx mam