Assembly Sample Equalization Using Fuzzy Granule Layers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inhomogeneous samples in product assembly processes with different styles pose challenges for machine learning algorithms due to subtle differences in function and strong internal data relevance, leading to inaccurate predictions and insufficient generalization ability.
Innovation Solution
An inhomogeneous sample equalization method and system that calculates similarity among samples, constructs a fuzzy compatibility matrix, clusters samples into granule layers, and performs equalization processing to identify an optimal granule layer for representative and accurate sampling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine learning models are trained on assembly process data of products with different styles, then the model can handle diverse product variants, but the inhomogeneity of samples reduces prediction accuracy and generalization ability
Solution Approach 1:
The patent segments the heterogeneous assembly process data into multiple style-specific datasets. Each style's assembly process is treated as a separate sample group, allowing the model to learn distinct patterns for each product variant while maintaining overall versatility. This segmentation resolves the contradiction by organizing diverse data into manageable, style-specific clusters that improve both accuracy and generalization.
Solution Approach 2:
The patent applies local quality by treating different product styles with differentiated processing approaches. Each style's assembly process data is analyzed and processed according to its specific characteristics rather than applying a uniform treatment. This allows the model to capture local patterns specific to each style while maintaining overall model performance across diverse variants.
2Adaptability or versatility
If assembly process data from multiple product styles is collected to improve model coverage, then more product variants can be handled, but the strong internal data relevance makes sample homogenization difficult
Solution Approach 1:
The patent segments the complex heterogeneous dataset into style-specific sub-datasets, making homogenization manageable at each segment level. By dividing the overall dataset according to product styles, the patent reduces the complexity of homogenization while maintaining comprehensive coverage of multiple product variants.
Solution Approach 2:
The patent introduces a new dimension of organization by categorizing data along the product style dimension. This dimensional approach allows the system to handle diverse product variants systematically, transforming the complex homogenization problem into a structured multi-dimensional data organization task that is more manageable.
3Productivity
If additional functions are designed for products to improve sales volume, then product variety increases, but the resulting inhomogeneous assembly processes reduce model generalization ability
Solution Approach 1:
The patent applies local quality by recognizing that different product configurations (with or without additional functions) have distinct assembly characteristics. Each product variant's assembly process is processed with consideration of its specific functional requirements, allowing the model to accurately capture and generalize from diverse product types while maintaining high prediction accuracy for each variant.
Data Source
AI summary
The disclosure discloses an inhomogeneous sample equalization method and system for a product assembly process. The method includes the following steps of: A: calculating a similarity among different samples; B: constructing a fuzzy compatibility matrix S for representing the similarity among all the samples, and constructing a fuzzy compatibility space X with different granule layers through the fuzzy compatibility matrix S; C: based on a granular calculating mode, screening out a granule layer with a maximum comprehensive value of an information increment and the similarity among the samples from the fuzzy compatible space X to serve as an optimal granule layer; and D: carrying out equalization processing on the sample of the optimal granule layer.