A novel prediction model based on long short-term memory optimised by dynamic evolutionary glowworm swarm optimisation for money laundering risk
by Pingfan Xia; Zhiwei Ni; Xuhui Zhu; Qizhi He; Qian Chen
International Journal of Bio-Inspired Computation (IJBIC), Vol. 19, No. 2, 2022

Abstract: The accurate prediction of money laundering risk is conducive to the risk prevention, and its task implies the characteristics of time series. Long short-term memory (LSTM) is widely developed in time series problems, but its parameters need to be optimised for delivering good predictive capacity. In this work, a novel bio-inspired algorithm, named dynamic evolutionary glowworm swarm optimisation (DEGSO), is designed. DEGSO employs adaptive step-size strategy, dynamic evolutionary mechanism, and directional mutation mechanism for improving the search performance. The money laundering risk is identified via a combination of DEGSO and LSTM (DEGSO-LSTM). DEGSO is wielded to optimise the main parameters of LSTM. Numerical results on four test functions indicate that DEGSO delivers better performance. Performance test results demonstrate that DEGSO-LSTM performs better than other state-of-the-art approaches. DEGSO-LSTM also attains satisfactory results in money laundering risk prediction, and its potential as a solution to financial fraud risk prediction.

Online publication date: Tue, 01-Mar-2022

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Bio-Inspired Computation (IJBIC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com