Risk Management Accelerator
Identify potentially risky activities in a high-frequency event stream using TIBCO's Risk Management Accelerator, using Machine Learning. Build supervised and/or unsupervised models and hot deploy these to the event processing platform. Score events in real-time and raise cases for investigation when potentially risky behaviour is detected.
TIBCO Spotfire® TIBCO® Enterprise Runtime for R (TERR™) TIBCO® Live Datamart TIBCO StreamBase® TIBCO ActiveMatrix® BPM
TIBCO ActiveMatrix BPM
TIBCO Live Datamart
TIBCO LiveView Web
TIBCO Spotfire Server
TIBCO Spotfire Analyst
TIBCO Spotfire Statistics Services
TIBCO Enterprise Message Service
TIBCO Component Exchange License
The accelerator is a demonstration of the Streaming Machine Learning end-to-end implementation. It starts with TIBCO Spotfire Analyst to guide data professionals through the process of developing both supervised and unsupervised machine learning models to detect risky behaviour in a given dataset. The Spotfire template is built like a machine learning tutorial and allows creating a model on your own data with less than 10 clicks. Therefore, the template adapts to the input data across different use cases and industries and can be used any time you want to separate a target group from the remainder. Example usecases include:
- in the FSI industry, credit scoring, collection scoring, anti-money laundering (AML), trade surveillance, insurance claim management, process management;
- in the manufacturing and transport industries, predictive maintenance based on IoT sensors, maverick lot detection, quality assurance;
- in CPG&R, cross-sell, up-sell, churn, and customer segmentation.
Models are developed using the R programming language and then hot deployed to TIBCO StreamBase for evaluation at runtime using TIBCO Enterprise Runtime for R.
The included demonstration detects potential fraud in a stream of financial transactions. When a transaction is scored, it will either pass or be flagged as similar to past fraud and/or plainly odd. When this occurs the event processing platform raises an alert and creates a case in TIBCO ActiveMatrix BPM that facilitates the human investigation of the potential risk in a totally audit-able way. Investigators are then provided with the entire context of the alert to help determine efficiently whether or not a transaction is actually fraud or not all within the fully functional case management environment. The results of investigations are tracked for continued model quality assurance and used within the next round of the model training, allowing the system to autonomously learn over time. The current accelerator therefore falls within the realm of prescriptive analytics decision support. Full automation and PMML usage are possible within the same technology.
The DevZone Forums are a traditional threaded discussion service subscribed to by Accelerator Developers, Practitioners, and Customers with a shared interest in the TIBCO Event Processing and Streaming Analytics platforms. Accelerators are provided as fast start templates and design pattern examples and are supported as delivered. For all questions concerning Accelerator use and implementation, please open a new discussion in the DevZone forum here.
There are currently no reviews for this content.
Why not be the first to review it - click here to login
Risk Management Accelerator
The Risk Management Accelerator illustrates a reference architecture to implement Streaming Machine Learning. We show how Predictive, Prescriptive and Streaming Analytics as well as Business Process Management can collaborate to support the full lifecycle of end-to-end risk detection. The definition of risk in this case is cross industry: the risk that a lot is maverick (manufacturing), the risk that a customer will churn (horizontal), the risk that a transaction is criminal (horizontal), the risk that a set of sensor readings herald a machine failure (manufacturing), the risk that a customer will say no (or yes) to an offering. in general, the accelerator is relevant whenever we want to use machine learning to single out a subset of our data from the rest. The accelerator includes a demonstration that shows an Anti-Money Laundering (AML) use case to detect potentially criminal transactions in a data stream.
This video explains why, within the financial crime use-case, the accelerator is important and how it works.
December 20 2017, release of Risk Management Accelerator 1.0.0
Companies have always had to make decisions. In an almost remote past, talent was enough to give a company competitive advantage - companies did best which employees had more talent to feel the market and formulate strategies accordingly. In the digital economy however this is increasingly insufficiently, not only because talent has always been scarce but because the availability of data allows us to mathematically support our opinions with evidence. And this again and again beats just talent.
Machine Learning is a branch of Artificial Intelligence that allows computers to detect patterns in data and automatically adapt those patterns change. When leveraging Machine Learning in risk scenarios, businesses can more efficiently spot where the riskiest data is today (e.g. those customers who are more likely to churn or those transactions that are more likely to be credit card fraud), but also understand the likely drivers of that risk. The net result allows them to draw policies for combating the risky scenarios effectively, e.g. tailored customer messaging that turns customers around for addressing the reasons for churn, or a revision of internal control procedures that address precisely the reason why crime almost went undetected. In turn, this maximizes customer satisfaction and retention as well as boosts business efficiency and productivity.
But machine learning tends to be considered complex by most business users, which inhibits its application across the breadth of organizations. TIBCO's accelerator allows a business field expert to get to a machine learning model within ten clicks, evaluate it, understand its drivers and apply it to today's data. The reusability of the platform across use-cases and the transparency of the end-to-end solution frees businesses from consultant dependency, effectively empowering the business to own its own success drivers. The front-end was created as a tutorial that demystifies the complexity and speaks to business experts who understand their KPIs.
Today’s dynamic risk detection systems need to be agile, scalable and intelligent.
- Agile to adapt to ever changing realities.
- Scalable to deal with ever increasing data volumes.
- Intelligent to detect increasingly sophisticated patterns, and also to reduce false positives in the alerting stage.
Coupled with the detection system, an important component is Case Management, which allows taking specific actions that target the risk. Taking an alert from when it is raised, through the investigation stage and if applicable on to the regulator requires a capable and integrated Case management tool, which can capture all the inputs from historical data, surveillance analyst notes and transaction log extracts.
Risk detection spans many different domains. Within financial services (FSI), some of these include Anti-money laundering (AML), Trade surveillance, Credit/Debit card monitoring, Health insurance fraud, Insurance fraud, Online operations. In manufacturing, risk can mean risk of producing defective items, or risk of machines working inefficiently, or risk of machines failing. In transport & logistics, there is the risk of losing luggage, or the risk of a car or boat needing repair or being deviated from its optimal course. In customer relations there is the risk of a customer feeling spammed.
In all cases what they have in common is a relatively high importance of transactions, at which any given time a certain percentage can be classified as risky and therefore deserving of special treatment. This percentage will vary greatly by market, industry, value, and other parameters. But, even if the content of the data differs greatly among them, the machine learning techniques can, at least initially, be widely similar.
The key to Streaming Machine Learning action systems is to maximize the number of cases that are identified, while minimizing the number of false detections (false positives) to come up with an optimal system both from an operations and financial standpoint.
Benefits and Business Value
The accelerator brings the following key benefits.
On the Predictive Analytics side:
- use your historic data to find patterns that represent risk,
- use TERR to create models that capture those patterns,
- do what-if analysis to set thresholds for model results so as to adjust your actions to the size of your team or budget,
- have your business users own this process from easy to use interfaces, without expensive consultancy,
- data discovery, machine learning model creation and real time deployment within a single tool.
On the Streaming Machine Learning side:
- use the latest best model defined by your business people to
- score every transaction in real time,
- separate risky ones and hand them to case management,
- notify user groups, sending not just alerts but also all their contextual info,
- keep model versioning,
- accelerate the adoption and rolling out of real time action: A graphical flow based development reduces complexity and increases collaboration; projects can span multiple assets, data and scenarios; with connectivity to over 150 options, including Bloomberg, BM&FBovespa, Currenex, EBS, FIX, FXall, Hotspot, Interactive Data, and Thomson Reuters, all your data feeds can be included.
On the action side in BPM:
- keep track of all alerts and all actions taken on them, such that everything is auditable at any time in the future,
- track on-going model quality by controlling false positive rates,
- integrate analytics dashboards that grab contextual, historical and reference data about alerts alongside information from the real-time processing,
- users can filter, slice and dice, zoom in and out of data in order to quickly determine if an alert needs further investigation or can be discarded as a false positive, which in turn allows,
- faster time to resolution,
- police the action - zoom out of each alert and verify that all members of the team are following all expected procedures, identify bottle-necks, manage workload,
- include the results of investigations in future iterations of the machine learning models.
The accelerator include a demonstration of fraud detection using a credit card transaction dataset that is analyzed against both a supervised and an unsupervised machine learning model to produce a score that indicates the risk's probability, and another score that show how the transaction deviates from what could be considered normal.
Other demo scenarios will vary from this. For example trade surveillance will differ greatly from insurance claims and credit card transactions. However the accelerator is intended to show a design pattern for detecting fraud using a combination of event processing and analytics rather than provide a series of concrete rules for doing the detection. The agility of the analytics and event processing platforms allow customers to modify the framework quickly and easily to adapt to other scenarios.
TIBCO software products and versions used:
|TIBCO ActiveMatrix BPM||4.2.0|
|TIBCO Live Datamart||7.7.1|
|TIBCO LiveView Web||1.3.1|
|TIBCO Spotfire Server||7.10.0|
|TIBCO Spotfire Analyst||7.10.1|
|TIBCO Spotfire Statistics Services||7.5.0 HF3|
|TIBCO Enterprise Message Service||8.3.0|