Dynamic Interaction Graph Feature Extraction for Sequence Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods face challenges in accurately expressing features of interaction objects due to the dynamic nature of interactions, which complicates the analysis and processing of interaction sequences, especially in capturing the time-order relationships and mutual impacts between interaction objects.
Innovation Solution
A dynamic interaction graph is constructed based on a sequence of interactions, where nodes represent objects and edges reflect their relationships over time, allowing for the extraction of feature vectors that comprehensively represent interaction objects by considering the impacts from other objects, using techniques like neural networks for feature extraction and prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional feature expression methods are used for interaction objects, then the implementation is simple, but the feature expression accuracy is insufficient due to the dynamic nature of interactions
Solution Approach 1:
The patent segments the interaction analysis into multiple components: interaction objects are divided into first and second objects, features are segmented into object features and interaction features, and the modeling process is divided into constructing interaction graphs and extracting features from sub-graphs. This segmentation allows accurate capture of dynamic interaction characteristics while managing complexity through modular processing.
Solution Approach 2:
The patent transitions from traditional flat feature representation to a multi-dimensional interaction graph structure. It introduces temporal dimensions by arranging interactions in time order, structural dimensions through graph nodes and edges, and relational dimensions by capturing interactions between objects. This dimensional expansion enables accurate feature expression of dynamic interactions.
2Measurement precision
If dynamic interaction graphs with comprehensive node relationships are constructed, then the feature extraction accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent extracts only the necessary sub-graph corresponding to the target object from the complete interaction graph, rather than processing the entire graph. This extraction focuses computational resources on relevant interactions and objects, improving feature extraction accuracy while reducing unnecessary computational overhead from unrelated parts of the interaction network.
Solution Approach 2:
The patent applies partial action by focusing on the specific sub-graph related to the target object rather than analyzing all interactions in the system. This selective approach provides sufficient information for accurate feature extraction without the excessive computational cost of processing the complete interaction graph, achieving the right balance between accuracy and efficiency.
3Reliability
If interaction sequences are analyzed to capture time-order relationships, then the understanding of interaction dynamics improves, but the processing difficulty increases
Solution Approach 1:
The patent performs preliminary action by pre-construction of the interaction graph organizing all interactions in time order before feature extraction. This pre-organization of interaction sequences with proper temporal ordering simplifies the subsequent feature extraction process, making it easier to capture time-order relationships without increasing processing difficulty during the actual analysis phase.
Data Source
AI summary
Computer-implemented methods, computer-implemented systems, and non-transitory, computer-readable media for processing interaction sequence data are disclosed. One computer-implemented method includes: obtaining a dynamic interaction graph is obtained, where the dynamic interaction graph is constructed based on a dynamic interaction sequence, including a plurality of interactions arranged in a chronological order, where each interaction includes two objects involved in the interaction and a time of the interaction. In the dynamic interaction graph, a sub-graph corresponding to a target node is determined, where nodes in the sub-graph comprise the target node and connection nodes connected to the target node through a predetermined amount of edges originating from the target node. A feature vector corresponding to the target node is determined based on a node feature of each of the nodes of the sub-graph and directions of edges of the sub-graph.