Java error on using k-NN Global Anomaly Detection with missing values
Hi, I am using RM Studio 8.2 and I have a process in which I am using k-NN Global Anomaly Detection to find anomaly scores for a set of data instances that have missing values in some of the attributes. In doing so, I get a Java error message "Process failed. Comparison method violates its general contract!". When I replace missing values before this operator, everything goes fine. However, this is not what I am after as I would like to find anomalies before treating missing values. Any ideas why I get that error? Thanks,
hello @b_ofoghi welcome to the community! Some quick requests so we can help you:
• Post your XML process here in this thread (see this post for instructions on How to Post on the Community)
• Attach your dataset if possible (use a fictionalized version if there are privacy concerns)
• Make sure you have all necessary extensions installed (see https://youtu.be/pjBqG3xtXx4)
I worked it out. The data set had some rows of full missing values (with no values in any attribute), once these are filtered out, the outlier detection operator works fine even if there are missing values scattered here and there.