choosing the best clustering algorithm depends on your data sparsity. But generally speaking K-nearest neighbors is a good choice for words clustering.

Thank you dear Tiba Zaki Abdulhameed for your quick replay. I am working on IoT non-spherical and unbalanced data for finding sensor fault detection.

Sivadi Balakrishna Firstly you got to check the assumptions of clustering models and test if they meet the criteria with your data. I would recommend to use Gaussian mixture model and make an assumption that every machine mode arrives from a Gaussian distribution. I have perform a lot of comparative studies of different models with IoT data and GMM has always proved to be reliable. Check out my profile for papers.

Thank you dear Nagdev Amruthnath for your massive explanation. I can look up your articles for understanding the literature and clustering algorithms.

you are saying it is unbalanced, so first apply a way to balance it such as smot or oversampling. Applying ML needs first to look at the data and do some preparations. This step will highly impact the results.

Thank you dear Tiba Zaki Abdulhameed . I will look up on suggested topics

dear,

try to look at K-means, its a clustering machine learning algorithms

thank you dear Bendaoud Fayssal

Use CFS clustering

Thank you dear Adinarayana Salina

Dear

use BIRCH algorithm

Thank you Santosh Satapathy

use incremental clustering algorithms

Dear,

use HMM

Thank you Lakshminarayana Kodavali