Clearing Message Prediction Models for Issuer Liquidity Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic payment networks face challenges in accurately predicting the timing and parameters of clearing messages, leading to confusion for account holders and liquidity issues for issuers due to latency and uncertainty in fund availability, as well as an inability to determine when to release pending charges.
Innovation Solution
A modelling platform using machine learning models is employed to predict clearing message parameters in real-time or near real-time by training on historical transaction data, enabling accurate forecasting of clearing message timing and amounts, and facilitating better fund management by issuers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If account issuers hold funds in reserve for expected clearing messages, then reliability of fund availability is improved, but liquidity is worsened
Solution Approach 1:
The system performs preliminary prediction of clearing message arrival times using machine learning models trained on historical transaction data. By predicting when clearing messages will arrive before they actually do, the system allows account issuers to release funds in advance while maintaining reliability through accurate predictions, thus improving liquidity without sacrificing fund availability assurance
Solution Approach 2:
The system implements a feedback mechanism where actual clearing message arrival data is continuously fed back into the machine learning models to refine and improve prediction accuracy over time. This feedback loop ensures that predictions become increasingly reliable, allowing the system to maintain high fund availability assurance while optimizing liquidity management through progressively accurate timing predictions
2Productivity
If clearing message latency is reduced, then productivity is improved, but measurement precision is worsened
Solution Approach 1:
The system performs preliminary analysis of authorization request data and historical patterns before clearing messages arrive, building predictive models that estimate clearing arrival times. This preliminary action enables the system to prepare and process clearing messages more efficiently without sacrificing prediction accuracy, as the models are trained on extensive historical data to maintain precision even as processing speeds increase
Solution Approach 2:
The system changes parameters of the machine learning models dynamically based on observed clearing patterns, transaction types, and time-of-day variations. By adjusting model parameters such as prediction time windows and confidence thresholds, the system can optimize for both faster clearing processing and maintained prediction accuracy across different operational conditions
3Measurement precision
If machine learning models are trained on historical data, then prediction accuracy is improved, but device complexity is worsened
Solution Approach 1:
The system introduces a specialised modelling platform as an intermediary component that handles all machine learning model training, validation, and deployment tasks. This intermediary layer separates the complex AI/ML functionality from the core payment processing system, allowing prediction accuracy to be improved through sophisticated model training while containing complexity in a dedicated module that can be updated independently without affecting the rest of the payment network infrastructure
Solution Approach 2:
The system segments the machine learning functionality into distinct modular components: data retrieval modules, model training modules, model validation modules, and prediction deployment modules. Each segment can be developed, tested, and maintained independently, reducing overall system complexity while enabling high prediction accuracy through specialized processing in each segment. This modular architecture allows the payment network to adopt advanced ML techniques without overwhelming system complexity
Data Source
AI summary
A modelling platform including at least one processor in communication with a memory device and a payment processor is provided. The at least one processor is configured to retrieve subsets of data from a transaction history database, derive training data sets from the subsets, apply model input data fields of each training data set as inputs to one or more machine learning models, and apply a machine learning algorithm to adjust parameters of the one or more machine learning models. The at least one processor is also configured to upload at least one trained machine learning model to an operational predictive model module, apply a stream of real-time authorization request messages as inputs to the at least one trained machine learning model, and transmit, to the payment processor in real-time or near real-time, values of at least one output obtained by applying the stream.