Automated Standing Settlement Instruction Selection via Ensemble ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual selection of standing settlement instructions (SSIs) is error-prone and fails to accurately reflect customers' settlement preferences, leading to incorrect recordings and subsequent trade corrections.
Innovation Solution
A machine learning-based system that combines a generic SSI sub-model and a customer-specific SSI sub-model to predict the most likely SSIs for a customer, using historic trade data and classification models like logistic regressions, decision trees, and neural networks to improve accuracy and automate the selection process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of standing settlement instructions is used, then customers can specify their preferences, but errors occur and accuracy decreases
Solution Approach 1:
The system enables automated self-service by using machine learning models to automatically select the most appropriate SSI for each trade based on historical data and trade attributes, eliminating manual intervention and reducing human error in SSI selection
Solution Approach 2:
The patent replaces the manual mechanical process of SSI selection with an automated electronic system using classification models (logistic regression, decision trees, random forests, neural networks) that analyze trade attributes and predict the most suitable SSI, thereby substituting human judgment with algorithmic decision-making
2Measurement precision
If support staff manually correct trades with incorrect SSIs, then trade accuracy can be improved, but productivity decreases and manual effort increases
Solution Approach 1:
The system performs preliminary action by training classification models on historical trade data beforehand, enabling these models to automatically and accurately select SSIs for new trades without requiring subsequent manual correction, thus preventing errors before they occur
Solution Approach 2:
The automated SSI selection system serves itself by using historical data to train models that independently make SSI selection decisions, eliminating the need for support staff intervention and freeing them to focus on higher-value tasks
3Ease of operation
If generic SSI reference data is maintained, then a baseline selection can be provided, but it fails to reflect individual customer preferences
Solution Approach 1:
The patent applies local quality by training separate classification models for different customers using their individual historical trade data, allowing each customer to have a customized model that reflects their specific preferences and trading patterns, rather than using a one-size-fits-all approach
Data Source
AI summary
Systems and methods for automated selection of standing settlement instructions are disclosed. According to one embodiment, in an information processing apparatus comprising at least one computer processor a method for automated selection of standing settlement instructions may include: (1) receiving, from a trading platform, historic trade data for a plurality of customers; (2) training a generic SSI sub-model with the historic trade data for the plurality of customers, resulting in a generic SSI prediction; (3) training a customer-specific SSI sub-model with the historic trade data for one of the customers, resulting in a customer-specific SSI prediction; (4) combining the generic SSI prediction and the customer-specific SSI prediction resulting in a final classification model; (5) receiving a trade to be executed for the customer; (6) applying the final classification model to the trade, resulting in a final SSI selection; and (7) executing the trade using the final SSI selection.