Signal Selection Device for Model Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Selecting appropriate data for model generation in data analysis is challenging due to the complexity of setting conditions, leading to potential inaccuracies in model accuracy.
Innovation Solution
A signal selection device that acquires candidate signals and calculates features based on their relevance to training labels, allowing for the selection of training signals for improved model accuracy in data analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of training data is performed based on predefined conditions, then data selection can be controlled, but the complexity of setting conditions increases and requires technical expertise
Solution Approach 1:
The system automatically selects training data by calculating feature values from candidate signals and automatically determining relevance to training labels without requiring manual condition setting. The selection is performed autonomously based on calculated features and pre-defined training labels, eliminating the need for complex manual condition configuration while maintaining high model accuracy.
Solution Approach 2:
The system changes the approach from manual condition-based selection to automated feature-based selection. By calculating feature values from candidate signals and using these features to determine relevance to training labels, the system transforms the selection process into a parameter-driven automated process that maintains precision without increasing operational complexity.
2Measurement precision
If manual selection of training data is performed, then data can be selected based on expertise, but the operation time and labor requirements increase
Solution Approach 1:
The system performs automated data selection by calculating feature values and automatically determining relevance to training labels. This self-service approach eliminates manual intervention, significantly reducing the time and labor required for data selection while maintaining the accuracy that would otherwise require expert judgment.
Solution Approach 2:
The system pre-calculates feature values from candidate signals before the selection process. By preparing feature calculations in advance and using them for automated relevance determination, the system reduces the overall time required for data selection operations while maintaining high model accuracy.
3Productivity
If inappropriate data is selected for model generation, then the model can be built quickly, but the model accuracy decreases
Solution Approach 1:
The system uses training labels as feedback to guide the selection process. By calculating feature values and comparing them against training labels to determine relevance, the system ensures that only appropriate data is selected for model generation. This feedback mechanism maintains both high productivity and high accuracy by automatically filtering out inappropriate data.
Solution Approach 2:
The system changes the selection criterion from manual condition-based to automated feature-based parameter comparison. By calculating feature values and using these parameters to determine relevance to training labels, the system ensures accurate data selection while maintaining quick model generation through automated processing.
Data Source
AI summary
A signal selection device (10) selects, from a plurality of candidate signals, a training signal for learning a model (20) usable for data analysis. The signal selection device (10) includes a first acquirer (11) that acquires the plurality of candidate signals and training labels associated with the plurality of candidate signals and being status values corresponding to results of analysis performed using the model, a feature calculator (13) that calculates one or more features for each of the plurality of candidate signals, and a selector (16) that selects the training signal from the plurality of candidate signals based on a degree of relevance between the one or more features and the training labels.