Big Data Machine Learning System Model Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current modeling methods, such as data modeling and simulation modeling, face limitations in accurately predicting system behavior when system configurations or rules change, and struggle with handling unexpected events due to the reliance on prior knowledge and data availability.
Innovation Solution
A computing system that integrates big data machine learning by using verified parameter values from machine learning on real-world data to build a system model, combining domain knowledge with machine learning algorithms to create a hypothetical model that can adapt to changes and predict system performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a machine learning-based model is built using big data, then prediction capability is improved, but the model cannot adapt when system configuration or operating rules change
Solution Approach 1:
The patent implements a dynamic modeling approach where the system model automatically adapts to changing conditions by continuously learning from new operational data. The model transitions from a static structure to a dynamic one that can adjust its parameters and behavior based on real-time system state changes, enabling it to maintain prediction accuracy despite configuration modifications.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously monitors actual system behavior, compares it with model predictions, and uses the discrepancies to refine and update the model. This closed-loop feedback allows the model to learn from actual system responses and adapt to changing operating rules, resolving the contradiction between initial prediction accuracy and long-term adaptability.
2Productivity
If data modeling method is used to represent correlation between datasets, then prediction can be made assuming system operates without change, but the method cannot handle sudden or changing situations
Solution Approach 1:
The patent transforms the static data modeling approach into a dynamic system that continuously updates its understanding of system behavior. By incorporating real-time data streaming and adaptive learning algorithms, the model maintains the speed of data-driven predictions while gaining the flexibility to handle sudden changes and evolving system conditions that static correlation models cannot accommodate.
3Reliability
If simulation modeling based on physical laws is used, then causal relationship can be expressed, but detailed system information must be available and model validation is required
Solution Approach 1:
The patent merges the strengths of both simulation modeling (causal relationship representation) and data-driven modeling (adaptive learning) into a hybrid approach. The system combines physical law-based causal models with machine learning algorithms that learn from actual operational data, allowing the model to maintain causal interpretability while automatically validating and refining itself through real-world observations, thereby reducing the manual validation burden.
4Reliability
If big data machine learning is applied to verify parameter values, then model reliability is improved, but computational resources and processing time are consumed
Solution Approach 1:
The patent applies partial machine learning verification rather than exhaustive validation of all parameters. The system selectively applies ML algorithms to verify only the most critical or uncertain parameters, using computational resources efficiently by focusing verification efforts where they provide the most value. This partial action approach maintains reliability improvement while significantly reducing overall computational consumption compared to comprehensive verification.
Data Source
AI summary
Provided is a computing system for implementing a system model using big data machine learning, which is intended to build a hypothetical model, calculate verified parameter values by performing machine learning on big data acquired from an actual system, and apply the verified parameter values to the hypothetical model. A system modeling method of completing a simulation model for a target system by causing a hypothetical model defined by acquiring knowledge about the target system to perform machine learning on big data acquired by running and observing the target system includes defining a hypothetical model for a target system by finding acquirable information related to the target system.