What are some good sources to learn fraud/anomaly detection in normal/time-series data?



I would like to know more on fraud/anomaly detection. I am looking for good source or survey article/book etc out there which will give me some preliminary idea of the area.

Any suggestion is greatly appreciated.



Posted 2015-04-17T19:14:49.843

Reputation: 1 047

I have answered a very similar question here: http://datascience.stackexchange.com/a/5200/2452. My related answer on Cross Validated might also be helpful: http://stats.stackexchange.com/a/142725/31372.

– Aleksandr Blekh – 2015-04-17T23:29:30.693

You're very welcome. – Aleksandr Blekh – 2015-04-20T21:36:50.910



Usually this is done as outliers analysis (fraud is an outlier vs normal usage). For this aspect, you can find more info in the data mining: concepts and techniques book, even if general purpose book.

I am convinced that learning this kind of basis is needed to understanding the domain specific methods.

Michael Hooreman

Posted 2015-04-17T19:14:49.843

Reputation: 713

You are welcome. Hoping this helps... – Michael Hooreman – 2015-04-21T07:54:48.467


Chandola et al's survey is probably the best and most-widely cited survey in the anomaly detection field.


Posted 2015-04-17T19:14:49.843

Reputation: 903