Machine Learning Models for Predictive Transaction Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for predicting future transactions in asset management, such as those used by registered investment advisors (RIAs), rely on less-than-accurate rules-based recommendations and lack advanced artificial intelligence and machine learning, resulting in lower confidence and fewer actionable insights for asset managers.
Innovation Solution
A predictive analysis system utilizing a server computing device that trains multiple machine learning models on historical transaction data to forecast future transactions by creating initial feature sets, determining target transaction variables, generating variable-specific feature sets, and executing trained models to generate predicted likelihood values for future transactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If rules-based recommendation systems are used for predicting future transactions, then the system implementation is simple, but the prediction accuracy and confidence are low
Solution Approach 1:
The patent replaces traditional rules-based mechanical recommendation systems with machine learning models that automatically learn patterns from historical transaction data. The system uses supervised learning algorithms to predict future transactions, substituting manual rule creation with automated data-driven models that achieve higher prediction accuracy while managing complexity through structured feature engineering and model selection processes.
2Reliability
If advanced machine learning techniques are implemented, then prediction accuracy and actionable insights improve, but system complexity and computational resources increase
Solution Approach 1:
The patent segments the prediction problem into multiple independent machine learning models, each trained on specific feature sets for different transaction types or time periods. This segmentation allows the system to achieve high reliability through specialized models while managing complexity by dividing the overall system into manageable components that can be developed, validated, and maintained independently.
Solution Approach 2:
The system employs parameter changes by training models on historical data with varying features, time periods, and transaction characteristics. The machine learning models automatically adjust parameters and weights based on the data, enabling the system to adapt to changing patterns in transaction behavior while maintaining high prediction confidence through continuous model training and validation.
3Productivity
If multiple machine learning models are trained on variable-specific feature sets, then the actionable insights and sales prediction capability improve, but the training time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-processing and organizing historical transaction data into structured feature sets before model training. The system performs feature engineering, data cleaning, and normalization in advance, creating ready-to-use feature matrices that can be quickly fed into multiple machine learning models. This preliminary preparation significantly reduces the actual training time while maintaining the ability to train multiple specialized models for different transaction scenarios.
Data Source
AI summary
Methods and apparatuses are described for predictive analysis of transaction data using machine learning. A server computing device trains a plurality of machine learning models using historical transaction data for a set of entities as input to predict a likelihood of future transaction activity for each of the entities, each machine learning model trained on a different target transaction variable. The server computing device executes each of the plurality of machine learning models to generate, for each entity, a predicted likelihood value for a future transaction associated with the entity and each of the target transaction variables. The server computing device transmits the predicted likelihood values for each entity to a remote computing device for display.