@misc{10481/68956, year = {2021}, month = {5}, url = {http://hdl.handle.net/10481/68956}, abstract = {The maintenance of cash levels under certain security thresholds is key for the health of the banking sector. In this paper, the monitoring process of branch network cash levels is performed using a single intelligent system which should provide an alert when there are cash shortages at any point of the network. Such an integral solution would provide a unified insight that guarantees that branches with similar cash features are secured as a whole. That is to say, a triggered alarm at a specific branch would indicate that attention must also be paid to similar (in-cash-feature) branches. The system also incorporates a (complementary) specific treatment for individual branches. The Early Warning System for securing cash levels presented in this paper (cash level EWS) is deliberately free of local demographic specifications, thereby overcoming the current lack of worldwide definitions for local demographics. This aspect would be particularly valuable for banking institutions with branch networks all over the world. A further benefit is the cost reductions that are a result of replacing several approaches with a single global one. Instead of local demographic parameters, a solid theoretical model based on Markov random fields (MRFs) has been developed. The use of MRFs means a reduction in the amount of information required. This would mean a higher processing speed as well as a significant reduction in the amount of storage capacity required. To the best of the author's knowledge, this is the first time that MRFs have been applied to cash monitoring.}, organization = {Spanish Ministry of Universities PID2019-103880RB-I00}, organization = {Junta de Andalucia SEJ340}, publisher = {John Wiley & Sons}, keywords = {Cash level EWS}, keywords = {Intelligent system}, keywords = {Local demographics free}, keywords = {Markov random fields}, keywords = {Securing cash levels}, title = {A novel intelligent system for securing cash levels using Markov random fields}, doi = {10.1002/int.22467}, author = {GarcĂ­a Cabello, Julia}, }