OREANDA-NEWS. “We expect trade monitoring/surveillance tools to catch up with trading systems and level the playing field for compliance officers,” says Sudhir Gupta. “These systems would be highly automated, keeping a 24/7 watch on trading activities. They would be capable of crunching very large data sets in real time, employing statistical techniques suitable for the segment.” Excerpts:

“A state-of-the-art trade monitoring/surveillance system should be implemented in a manner that provides a complete view of the firm’s trading activity, i.e. across different lines of business and geographies. Currently, most banks have siloed environments in which these tools work.

As banks become more global, the trade monitoring/surveillance tools should have the ability to analyze traders’ activities on positions held in different asset classes and in all markets accessible to the trader. The trade monitoring/surveillance system needs to combine historical and real time, structured and unstructured data with predictive analysis tools to detect and flag events in real time. There should be a capability to crunch large data sets, potentially through the use of ‘Big Data’ technologies. Finally, systems would need to have dynamically evolving rules engines that can keep pace with the changes in regulations.

There are new and emerging techniques such as hybrid models of distance-based clustering and dynamic Bayesian network. The overall monthly activities of a trader establish a defined pattern. This pattern closely matches traders of similar characteristics and can be captured in the form of a probabilistic temporal model. Bayesian network may flag a set of trades if they fit into the context of the particular trader’s trading pattern.”​