Additive Data-Based Partial Model Chain for Local Effect Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data-based function models, such as Gaussian process models, struggle to accurately represent local effects and incorporate new training data without increasing measurement noise and modeling error, especially when additional training data records have insufficient measuring points or overlapping data points with existing records.
Innovation Solution
The method involves creating an additive chain of data-based partial models, where each additional partial model specifies an additive deviation from the initial model, allowing for effective representation of local effects without impairing the initial model, using a separate model calculation unit to simplify calculations and accelerate function value determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If additional training data records are merged into an existing data-based function model, then the model can incorporate more information, but measurement noise and modeling error increase when additional measuring points have insufficient weight or conflict with existing data
Solution Approach 1:
The patent segments the training data into multiple training data records processed sequentially. Each record is handled as a separate segment that contributes to the overall model through controlled integration, allowing the model to adapt to new data while maintaining stability from previous records.
Solution Approach 2:
The patent modifies hyperparameters during the sequential processing of training data records. By adapting hyperparameters based on the current training record being processed, the model can adjust its behavior to accommodate new data while controlling the impact on measurement noise and modeling error.
2Productivity
If a separate model calculation unit is added to calculate data-based models, then calculation speed is improved, but device complexity increases
Solution Approach 1:
The patent combines the model calculation functionality with the existing control unit structure. The separate model calculation unit is integrated into the control unit, allowing it to operate as a coordinated system that leverages existing hardware resources while providing accelerated calculation capabilities.
Data Source
AI summary
A method for generating a data-based function model includes: providing a first data-based partial model ascertained from a first training data record; providing at least one additional training data record; and performing the following steps for the at least one additional training data record: ascertaining a difference training data record having training data which correspond to the differences between the output values of the relevant additional training data record and the function value of the sum of the partial function values (ffirst_partial_model(x) fsecond_partial_model(x)) of the first data-based partial model and previously ascertained data-based partial model(s) at each of the measuring points of the relevant training data record; ascertaining an additional data-based partial model from the difference training data record; and forming a sum (f(x)) from the first and the additional data-based partial models.