Annotated Sample Selection for Cross-Field Emotion Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high labor and time costs associated with manual data annotation for emotion analysis systems, particularly when adapting to new fields, hinder the efficient deployment of deep learning technologies.
Innovation Solution
A method for selecting annotated samples by determining first and second attributes of sample characteristics in source and target field sample sets, allowing for the identification of target annotated samples to train classification models for emotion polarity analysis, thereby reducing manual annotation costs and enabling cross-field emotion analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual annotation is performed for each field to ensure model accuracy, then the reliability of emotion analysis is improved, but the labor cost and time cost increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-processing the source field sample set to extract sample characteristics and determine their attributes before cross-field selection. This preliminary extraction and attribute determination enables efficient reuse of annotated data across different fields without requiring manual annotation for each new field, thus reducing annotation time while maintaining model accuracy.
Solution Approach 2:
The patent implements copying by selecting annotated samples from a source field and reusing them for target field emotion analysis models. Instead of creating new annotated data for each field, the system copies relevant annotated samples from the source field based on characteristic attribute matching, significantly reducing the time and labor required for data annotation while preserving annotation quality.
2Reliability
If manual annotation is performed for each field to ensure model accuracy, then the reliability of emotion analysis is improved, but the labor cost increases significantly
Solution Approach 1:
The patent applies universality by creating a multi-functional system that can serve multiple fields using a single source field sample set. The extracted sample characteristics and their attributes enable the same annotated data to be universally applied across different target fields, eliminating the need for separate manual annotation processes for each field and thereby reducing labor costs while maintaining analysis reliability.
Solution Approach 2:
The patent implements copying by reusing annotated samples from the source field across multiple target fields. By copying relevant annotated data based on characteristic attribute matching rather than creating new annotations for each field, the system significantly reduces annotation labor costs while preserving the quality and reliability of emotion analysis.
3Loss of time
If cross-field sample selection is performed without attribute determination to reduce annotation costs, then the labor cost and time cost are reduced, but the accuracy of emotion analysis decreases
Solution Approach 1:
The patent applies local quality by determining specific attributes of sample characteristics at local levels (source field and target field separately) and then matching them for cross-field selection. This localized attribute determination ensures that each field's specific characteristics are properly identified and matched, maintaining high accuracy in emotion polarity determination while enabling efficient cross-field sample reuse to reduce annotation time.
Solution Approach 2:
The patent implements parameter changes by transforming the source field sample characteristics into a standardized attribute representation that can be compared across different fields. By changing the parameters to a common attribute framework (extracting and comparing characteristic attributes), the system enables accurate cross-field sample selection that maintains emotion analysis precision while reducing annotation time costs.
Data Source
AI summary
The present disclosure provides a method for selecting an annotated sample. The method includes: determining a first attribute and a second attribute of a sample characteristic; in which the first attribute is a characteristic attribute of the sample characteristic in a source field sample set, and the second attribute is a characteristic attribute of the sample characteristic in a target field sample set; and determining a target annotated sample from a plurality of candidate annotated samples of the source field sample set according to the first attribute and the second attribute; in which the target annotated sample is configured to train a classification model, the classification model includes a model for determining an emotion polarity by analyzing an input sample to be classified.