# Clustering based on threshold values.

11 views (last 30 days)
Silpa K on 4 Feb 2020
Commented: Silpa K on 5 Feb 2020
I have a dataset ,attached here. I need to cluster the data based on some threshold value that chhosen randomly from the data set.How can I cluster a data based on a point or a set of points. Please help me.

Hiro Yoshino on 4 Feb 2020
Is this a question about how to classify the data based on threshold?
if condition
"classification description"
else
"classification description"
end
can be a solution.

#### 1 Comment

Silpa K on 5 Feb 2020
I need to cluster it based on a distance measures.

Hiro Yoshino on 5 Feb 2020

Silpa K on 5 Feb 2020
Thank you sir.
My question is if I select some points from the dataset(suppose 2,1,0.2 etc). Based on this points how can I cluster. Cluster the different dense regions based on this points. Taking the points and cluster a point near to that point then we will get a cluster that contain 2 elements and clustering another point that near to the that cluster repeating this process.
Hiro Yoshino on 5 Feb 2020
Do you want to use fixed some points as centroids of the clusters?
OK, you've got an algorithm?
Then, is your question how to put the algorithm into practice in MATLAB?
Your algorithm sounds very straightfoward. It seems it is just a matter of how to write it in MATLAB. Am I right?
If so, you should start from basics: https://matlabacademy.mathworks.com/
MATLAB onramp is a good fit for you.

Silpa K on 5 Feb 2020
Thank you sir.
Please give me a suggestion that which clustering technique can I use to fixed som epoints as centroids.

Hiro Yoshino on 5 Feb 2020
As far as I know, the centroids of clusters are determined by the corresponding clusters. So it is uncommon to fix the centroids before finding the clusters.
I suppose you can write by yourself. It sounds very easy.
Algorithm as follows:
1. calculate the distances from all the centroids
2. classify the data into the nearest cluster
3. loop 1 and 2 till end
Silpa K on 5 Feb 2020
Is there is any matlab code or technique available to cluster the nearest elements of the point.