Hi,
As per my understanding, you are interested in deleting all the outliers obtained from DBSCAN clustering in MATLAB.
In order to do this, you can follow these steps:
- The MATLAB dbscan function returns an array idx, containing the value that represents the cluster (cluster label) for each index or point in the input dataset.
idx = dbscan(X,epsilon,minpts)
- The points which are outliers may be represented by -1 label, and you can easily locate those with the find() function.
outlierIndices = find(idx == -1);
- Finally you can remove the outliers in the dataset like this.
newData = dataset;
newData(outlierIndices, :) = [];
Here are some documentation links that you might want to follow:
Thank you.
