Trial Production Condition Proposal System Using Weighted Regression Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In material development using Material Informatics, the accuracy of estimating trial production conditions is compromised by the small number of learning data samples and biases in their distribution, leading to insufficient estimate accuracy.
Innovation Solution
A trial production condition proposal system that constructs regression models by calculating a weight reference for measured characteristics data and performing weighting based on this reference, enhancing the accuracy of proposed trial production conditions even with limited or biased data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the number of learning data samples is small, then the complexity of data collection is reduced, but the estimate accuracy of trial production conditions deteriorates
Solution Approach 1:
The patent applies parameter changes by transforming the learning data through weighting operations. It calculates importance weights for each sample based on distance metrics in the feature space, then applies these weights during model training. This transforms the original uniform-weighted small dataset into an effectively enhanced dataset where certain samples contribute more to the learning process, thereby improving estimate accuracy without collecting more data.
Solution Approach 2:
The patent creates a composite data structure by combining multiple weighted samples into an enhanced learning dataset. It synthesizes a new dataset where each original sample is replicated according to its calculated importance weight, creating a composite structure that better represents the underlying distribution. This composite approach allows the model to learn from limited data more effectively.
2Ease of manufacture
If the learning data distribution is biased, then the data collection process is simplified, but the estimate accuracy of trial production conditions deteriorates
Solution Approach 1:
The patent addresses distribution bias by dynamically changing the weight parameters of biased samples. It calculates importance weights based on distance to query points in the feature space, automatically down-weighting biased or redundant samples and up-weighting informative ones. This parameter transformation corrects the distribution bias without requiring complex re-sampling or data collection processes.
Solution Approach 2:
The patent implements feedback by using the query point information to dynamically determine sample weights. The importance of each training sample is evaluated based on its relevance to the specific query, creating a feedback loop where the model adapts to the data distribution characteristics. This feedback mechanism automatically corrects biases by emphasizing samples that are more informative for the given query.
3Loss of information
If the number of explanatory variables is large, then the information content is increased, but the difficulty of constructing accurate machine learning models increases
Solution Approach 1:
The patent applies local quality by treating different regions of the feature space differently through localized weighting. Instead of uniform processing, it calculates weights based on local distance metrics between samples and query points. This allows the model to focus computational resources on locally relevant features and samples, reducing the effective complexity while preserving important information.
Solution Approach 2:
The patent segments the large set of explanatory variables and samples by grouping them according to their importance weights. It divides the feature space into regions of high and low importance, processing weighted samples separately. This segmentation reduces the computational burden by focusing on the most informative subsets of the large-dimensional data.
Data Source
AI summary
A trial production condition proposal system 1 is a system for proposing a trial production condition(s) for a material to a material developer and includes a regression model construction processing unit 112 and a trial production condition proposal processing unit 113. The regression model construction processing unit 112 executes regression model construction processing on measured characteristics data indicating an actual measurement result of characteristics of the material. The trial production condition proposal processing unit 113 searches for an optimum trial production condition for the material by using the constructed regression model and executes trial production condition proposal processing based on a search result. The regression model construction processing includes: processing for calculating a weight reference which is a reference for weighting on the measured characteristics data; and processing for performing weighting on the measured characteristics data based on the calculated weight reference.