Yue Zhao

icon indicating a website link https://viterbi-web.usc.edu/~yzhao010/

icon company University of Southern California icon location Los Angeles, CA, USA icon location I HAVE No Twitter/X account! Assistant Professor of Computer Science @ University of Southern California (USC); Ph.D. @ CMU. Anomaly/Outlier/OOD Detection

Results 138 comments of Yue Zhao

Is there any possbility that a ordinary supervised model performs better than a outlier algorithm in this task?

A rule of thumb, if you have labels, using supervised models is preferred even for anomaly detection. Charu Aggarwal-- Outlier Analysis--Second Edition--Page 26 ![image](https://user-images.githubusercontent.com/15079146/87556860-290eb080-c685-11ea-8ff5-8aedcdad2052.png)

Autoencoder is giving segmentation fault.

Please share the screenshot...I am afraid that this is not about PyOD as it is the first time we heard this. If abod, knn, and other non deep learning models...

Parallelize the fit and decision_function methods of FeatureBagging

Sorry for being late. this will be the next PR to be in. It looks nice to me. Just need to run some test locally :) hang in there. Thanks!

Add optimal bin choice to LODA

noted and thanks for the advice

Add optimal bin choice to LODA

that will be great for LODA. Help appreciated!

ECOD behavior on new data

> Thanks for this analysis. Very helpful. I also agree that in the ideal case test samples should be somehow independent while predicting. So whether train+test or train+a single test...

knn metric cosine ValueError: Unrecognized metric 'cosine'

Thanks for the note. PR is welcomed :)

Locally Selective Combination (LSCP) ValueError: x and y must have length at least 2

Thanks for the note. I will say just rerun it. I tried to reproduce on my end and it went through. if possible, could you share me the code and...

Locally Selective Combination (LSCP) ValueError: x and y must have length at least 2

Problem acknowledged. Looks like there are some update on the dependent libraries.

Locally Selective Combination (LSCP) ValueError: x and y must have length at least 2

as shown https://github.com/yzhao062/pyod/pull/192, just change line 74 to `random_state = 42` There are some underlying changes in numpy and sklearn, which cause the issue. A PR will be merged soon.