Causal Prediction Model for Product Quality Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction systems struggle to accurately determine how to adjust explanatory variables to achieve desired product quality values, often requiring high calculation costs and lacking clarity in causality between variable changes.
Innovation Solution
A prediction apparatus that generates a prediction model based on causality information, using process data from a production facility to determine the positive/negative variation direction of response variables in response to changes in explanatory variables, thereby reflecting the causality between variable variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a prediction model is generated using traditional methods without causality information, then the model can be built quickly, but the model cannot accurately determine how to adjust explanatory variables to achieve desired product quality values
Solution Approach 1:
The system performs preliminary action by pre-defining causality information that specifies the variation direction (positive or negative) of response variables when explanatory variables change. This causality information is established before the prediction model is constructed, allowing the model to inherit these directional relationships and directly determine how to adjust explanatory variables to achieve desired quality values without requiring complex post-processing or iterative optimization.
2Reliability
If a method of changing combination of explanatory variables and repeating simulation is used to solve inverse problems, then the desired result can be obtained, but the calculation cost becomes very high
Solution Approach 1:
The system implements feedback by incorporating causality information that directly links explanatory variables to response variables with specified variation directions. When a desired quality value is targetted, the system can immediately determine the required adjustment direction of explanatory variables based on the pre-established causality relationships, eliminating the need for repeated simulations and significantly reducing calculation time while maintaining control accuracy.
3Ease of operation
If traditional prediction models are used without causality constraints, then the model generation is simpler, but it is not clear how to change explanatory variable values to bring estimated values closer to desired values
Solution Approach 1:
The system applies parameter changes by embedding causality information that defines the variation direction (positive or negative) of response variables corresponding to changes in explanatory variables. This transforms the prediction model from a black-box estimator into an interpretable system where the relationship between input changes and output responses is explicitly defined, making it straightforward to determine how to adjust explanatory variables to achieve desired quality values.
Data Source
AI summary
A regression model reflecting causality between variation of an explanatory variable and variation of a response variable is constructed. A prediction apparatus predicts characteristic values of a product by using process data obtained from a production facility. The prediction apparatus includes a process data acquisition unit that reads the process data from a storage device that stores the process data obtained from the production facility, and a prediction model generation unit that generates a prediction model on the basis of causality information that defines a combination of first process data and second process data or a value corresponding to the second process data. The first process data and the second process data or the value corresponding to the second process data are included in the read process data. The first process data is used as a predetermined explanatory variable. The second process data or the value corresponding to the second process data is used as response variable. The prediction model has learned features of the process data obtained from the production facility. The prediction model generation unit generates the prediction model and determines a positive/negative variation direction of the response variable in accordance with a positive/negative variation direction of the predetermined explanatory variable.