Communication Completion Prediction Using Network Trigger Rules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems lack mechanisms to determine hard disk speed of receiving computers, network routes, and communication loads, making it difficult to predict communication completion times accurately.
Innovation Solution
A two-tiered machine learning architecture is used to independently determine probabilities of various factors affecting communication completion, aggregating these probabilities to select a rule set for predicting when communications will be completed, thereby overcoming uncertainty and bias.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems are used to predict communication completion, then the system structure remains simple, but prediction accuracy deteriorates due to inability to determine hard disk speed, network route, and communication load
Solution Approach 1:
The patent segments the prediction system into multiple independent components: a machine learning model for predicting communication completion time, a separate module for determining hard disk speed, another for identifying network routes, and a third for calculating communication load. Each component processes specific factors independently and feeds into the overall prediction, allowing high accuracy without requiring a single monolithic complex system.
Solution Approach 2:
The patent introduces an intermediary data processing layer that collects and processes various network parameters (hard disk speed, network route, communication load) as intermediate inputs to the machine learning prediction model. This intermediary layer aggregates data from multiple sources and transforms them into meaningful features for prediction, bridging the gap between raw data and accurate prediction.
2Measurement precision
If multiple factors (hard disk speed, network route, communication load) are considered in prediction, then prediction accuracy improves, but difficulty of detecting and measuring these factors increases
Solution Approach 1:
The patent implements self-service mechanisms where the system automatically detects and measures network parameters without external intervention. The hard disk speed is determined by monitoring disk I/O operations, network routes are identified through packet tracing, and communication load is calculated by analyzing traffic patterns. The system serves its own measurement needs by incorporating these detection capabilities directly into the prediction infrastructure.
Solution Approach 2:
The patent employs feedback mechanisms where the system continuously monitors actual network conditions and uses this information to refine future predictions. By comparing predicted communication completion times with actual outcomes, the system learns from past performance and adjusts its measurement and prediction models, making it progressively better at detecting and measuring network factors.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A system for processing performance prediction decisions includes one or more processors configured to execute one or more program modules. The modules are configured to receive, via the one or more processors, a prediction for an account at a prediction timestamp. The modules are also configured to identify, via the one or more processors, a prediction rule using attributes from the prediction. Responsive to the prediction rule having a network trigger associated therewith, the modules are configured to retrieve, via the one or more processors, a network trigger time associated with the network trigger, compare, via the one or more processors, the prediction timestamp to the network trigger time, and apply, via the one or more processors, a prediction decision based on the comparison of the prediction timestamp and the network trigger time. Applying the prediction decision includes determining a confidence level that a communication associated with the prediction will occur by a given time.