Sticker Recommendation System Using Dialogue Context Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for recommending stickers in social network services and instant messengers often fail to consider the context of a dialogue, leading to inappropriate emoticon or sticker suggestions, as they rely solely on keyword matching and unitary utterance analysis, which limits their ability to accurately reflect the emotional state and situation.
Innovation Solution
A method that analyzes a series of continuous utterances to determine the context and emotional state, using surface and emotion analysis to extract dialogue context factors, and recommends stickers based on these factors, considering the relationship with the other party and personal preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If keyword matching is used for sticker recommendation, then the system is simple and fast, but the recommendation accuracy and contextual appropriateness deteriorate
Solution Approach 1:
The patent segments the dialogue analysis into multiple independent modules: surface analysis (morphological, syntactic, semantic), emotion analysis (emotion type detection, emotion strength detection), and context analysis. Each module processes specific aspects of the utterance independently, allowing parallel processing while maintaining comprehensive analysis. This segmentation enables the system to achieve high recommendation accuracy without sacrificing processing efficiency.
Solution Approach 2:
The patent performs preliminary surface analysis and emotion analysis on the current utterance before generating sticker recommendations. By pre-processing the utterance to extract emotion types, emotion strengths, and contextual features, the system prepares the necessary information in advance, enabling fast and accurate recommendation generation without requiring complex real-time computation during the recommendation phase.
2Productivity
If only current utterance is analyzed, then the processing is simple and fast, but the contextual understanding deteriorates
Solution Approach 1:
The patent extends the analysis from a single utterance dimension to a multi-utterance dialogue dimension by incorporating historical dialogue context. The system analyzes not only the current utterance but also previous utterances in the dialogue, extracting emotion history and contextual patterns across multiple turns. This dimensional extension allows the system to capture ironic usage and contextual nuances while maintaining efficient processing through structured context representation.
3Device complexity
If emotion analysis is performed by unit of utterance, then the analysis is simple, but the recommendation considering context becomes impossible
Solution Approach 1:
The patent merges the analysis of multiple utterances by combining surface analysis results, emotion analysis results, and context analysis results into a unified recommendation framework. The system integrates emotion types, emotion strengths, and contextual features from both current and previous utterances to generate contextually appropriate sticker recommendations. This merging approach enables comprehensive contextual understanding while maintaining manageable analysis complexity through structured integration of multiple analysis layers.
Data Source
AI summary
Provided is a method of recommending a sticker through an emotion analysis. The method of recommending a sticker through an emotion analysis, include: by a server, performing a surface analysis on the last utterance between the first user terminal and the second user terminal; performing an emotion analysis on the last utterance using a result of the surface analysis; extracting a dialog context factor including a surface analysis result and an emotion analysis result on a certain number of continuous utterances including the last utterance between the first user terminal and the second user terminal; selecting a sticker to be recommended to the first user using the dialog context factor; and providing the selected sticker for the first user terminal.