Propensity Models for Predicting Data Calls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional temporal and time series forecasting methods are inaccurate due to variability in temporal data, leading to significant latency issues in predicting external data calls, which affects computing service quality and user experience.
Innovation Solution
Implementing automated sequence-based propensity models using deep neural networks (DNNs) with architectures like LSTM and GRU to forecast future external data calls by encoding past events as sequences, allowing for preprocessing and caching of data to reduce latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional temporal and time series forecasting methods are used, then the system is simpler to implement, but the prediction accuracy deteriorates due to variability in temporal data
Solution Approach 1:
The patent replaces conventional statistical forecasting methods with deep neural network models (LSTM, GRU, Transformer) that use encoded sequences of past events to predict future data calls. This substitution enables the system to capture complex temporal patterns and relationships that conventional methods miss, significantly improving prediction accuracy while managing complexity through automated sequence encoding and propagation mechanisms.
2Reliability
If external data calls are made to improve service quality, then service performance improves, but processing latency increases
Solution Approach 1:
The patent implements preliminary action by using propensity models to predict which external data calls are likely to be needed in the future, allowing the system to prefetch and cache this data in advance. This enables the system to make service quality improvements through external data calls without incurring real-time latency penalties, as the data is already prepared and available when needed.
Data Source
AI summary
There are provided systems and methods for reducing latency through propensity models that predict data calls. A service provider, such as an electronic transaction processor for digital transactions, may provide computing services to users including those for electronic transaction processing. In order to provide sequence-based forecasting of computing events and processing requests for users, accounts, and/or activities associated with the service provider, the service provider may provide a machine learning model, such as a deep neural network, that predicted potential occurrences and likelihoods of computing events occurring at future times. When predicting the events, the service provider's machine learning predictive framework may further predict data calls required to be executed to retrieve data needed for processing during the events. These predicted calls may then be batched together into a batch processing job, which may be executed to retrieve the data prior to the predicted events.