Maximum likelihood continuity mapping for fraud detection Page: 4 of 14
This report is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
Maximum likelihood continuity mapping for fraud detection
Los Alamos National Lab
Los Alamos, NM 87545
We describe a novel time-series analysis technique called maximum likelihood continuity
mapping (MALCOM), and focus on one application of MALCOM: detecting fraud in medical
insurance claims. Given a training data set composed of typical sequences, MALCOM creates a
stochastic model of sequence generation, called a continuity map (CM). A CM maximizes the
probability of sequences in the training set given the model constraints. CM's can be used to
estimate the likelihood of sequences not found in the training set, enabling anomaly detection
and sequence prediction -- important aspects of data mining. Since MALCOM can be used on
sequences of categorical data (e.g. sequences of words) as well as real valued data, MALCOM is
also a potential replacement for database search tools such as N-gram analysis. In a recent
experiment, MALCOM was used to evaluate the likelihood of patient medical histories, where
"medical history" is used to mean the sequence of medical procedures performed on a patient.
Physicians whose patients had anomalous medical histories (according to MALCOM) were
evaluated for fraud by an independent agency. Of the small sample (12 physicians) that has been
evaluated, 92% have been determined fraudulent or abusive. Despite the small sample, these
results are encouraging.
keywords: anomaly, fraud, time-series, continuity map
Here’s what’s next.
This report can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Report.
Hogden, J. Maximum likelihood continuity mapping for fraud detection, report, May 1, 1997; New Mexico. (https://digital.library.unt.edu/ark:/67531/metadc688007/m1/4/: accessed May 26, 2019), University of North Texas Libraries, Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.