Reduced Training Set for Electronics Feature Labeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing the relationship between features and labels in electronics are inefficient due to high computational load and slowness, making it difficult to monitor and predict quality attributes of electronic items effectively.
Innovation Solution
A method involving a processing unit and memory that aggregates data into a reduced training set by grouping similar feature values and label values, reducing the data size while maintaining accuracy, and using decision trees for classification, thereby reducing computational complexity and enabling quicker model building and prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If current methods are used to analyze the relationship between features and labels, then model accuracy can be maintained, but computational load is high and processing speed is slow
Solution Approach 1:
The patent segments the large training dataset into multiple smaller batches or chunks that can be processed independently and in parallel. This segmentation reduces the computational load on individual processing units while maintaining the ability to process the complete dataset through aggregation of results from multiple batches.
Solution Approach 2:
The patent introduces a new dimension of processing by implementing a two-stage approach: first processing data in reduced batches to identify candidate relationships, then validating and refining these relationships across the complete dataset. This dimensional approach to data processing enables faster initial analysis while preserving accuracy through subsequent verification.
2Measurement precision
If the complete training set is processed, then model accuracy is maintained, but computation time increases significantly
Solution Approach 1:
The patent performs preliminary processing of the training data by creating aggregated representations and feature summaries before the main analysis. This preliminary action reduces the complexity of subsequent processing while preserving the essential information needed for accurate model training, thereby reducing overall computation time without sacrificing accuracy.
Solution Approach 2:
The patent implements continuous processing where results from batch processing are aggregated and used to inform subsequent processing steps. This continuity ensures that no useful information is lost during the divided processing, maintaining model accuracy while enabling parallel execution that reduces total computation time.
3Device complexity
If data is aggregated into reduced training sets, then computational complexity is reduced, but data representation fidelity may be compromised
Solution Approach 1:
The patent merges multiple data points into aggregated representations that capture the essential statistical properties of the underlying data distribution. By combining data in a way that preserves mean, variance, and other key characteristics, the aggregation reduces computational complexity while maintaining sufficient fidelity for accurate model training.
Solution Approach 2:
The patent transforms the raw training data into aggregated parameter representations (such as mean values, standard deviations, and frequency distributions) that capture the essential information in a compressed form. This parameter transformation reduces data dimensionality and computational requirements while preserving the statistical relationships needed for accurate analysis.
Data Source
AI summary
A method comprising, by a processing unit and a memory: obtaining a training set of data; dividing sets of data into a plurality of groups, wherein all sets of data for which feature values meet at least one similarity criterion, are in the same group, storing in a reduced training set of data, for each group, at least one aggregated set of data, wherein, for a plurality of the groups, a number of aggregated sets of data is less than a number of the sets of data of the group, wherein the reduced training set of data is suitable to be used in a classification algorithm for determining a relationship between the at least one label and the features of the electronic items, thereby reducing computation complexity when processing the reduced training set of data, compared to processing the training set of data.