Negative Example Availability Deciding Apparatus for Training Data Reuse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Training data used as negative examples for one classifier cannot be reused as negative examples for another classifier without excluding data that should be handled as positive examples, due to mixed association with various items lacking labels.
Innovation Solution
A negative example availability deciding apparatus that generates two classifiers, one trained with target data as a positive example and another as a negative example, to determine classification accuracy and decide the reuse of target data as negative example training data for another classifier.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If training data is reused as negative examples for multiple classifiers, then productivity is improved, but classification accuracy deteriorates when the data should be handled as positive examples
Solution Approach 1:
The system performs preliminary classification using a first classifier trained with target data as positive examples before reusing the data as negative examples. This preliminary action identifies which data items should be excluded, preventing accuracy deterioration while enabling reuse of suitable data for improving productivity
Solution Approach 2:
The training data is segmented into two groups: data that can be reused as negative examples and data that should be excluded. The availability deciding section divides the target data based on classification results, allowing selective reuse that maintains accuracy while improving efficiency
2Productivity
If all target data is used as negative examples without exclusion, then productivity is improved, but measurement precision deteriorates due to inability to identify data requiring exclusion
Solution Approach 1:
A third classifier serves as an intermediary tool to evaluate target data and identify items that should be excluded from negative example training. This intermediary classification process enables precise identification without manually reviewing each data item, maintaining both productivity and measurement precision
3Reliability
If manual review of target data is performed to identify data for exclusion, then classification accuracy is maintained, but productivity deteriorates
Solution Approach 1:
The system uses automated classification processes where the first and third classifiers automatically evaluate target data and identify items for exclusion. This self-service approach replaces manual review, maintaining classification accuracy through algorithmic evaluation while dramatically improving productivity by eliminating time-consuming human intervention
Data Source
AI summary
There are provided a negative example availability deciding apparatus, a negative example availability deciding method, and a program that enable reuse of training data. A positive example evaluation training section generates a positive example evaluation classifier trained with target data as a positive example. A negative example evaluation training section generates a negative example evaluation classifier trained with target data as a negative example. An availability deciding section determines the classification accuracy of the positive example evaluation classifier and the classification accuracy of the negative example evaluation classifier by using evaluation data. The availability deciding section decides, on the basis of the classification accuracy of the positive example evaluation classifier and the classification accuracy of the negative example evaluation classifier, whether or not to use target data as negative example training data for training another classifier.