Graphical Representation of Frame Instances for Opinion Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for an automated system to process and summarize vast amounts of opinion data from the internet, which is not efficiently harvested or utilized beyond its original source, particularly for marketing and brand management purposes.
Innovation Solution
A Frame-Based Search Engine (FBSE) system that produces frame instance data by applying frame extraction rules to natural language corpora, enabling graphical representations such as Instance Graphs and Instance Plots to visualize preferences between items, using a Preference Frame with roles like 'Item' and 'Preferred Item' to analyze consumer preferences across categories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If frame extraction rules are applied to process vast amounts of opinion data, then the ability to summarize and visualize consumer preferences is improved, but the system complexity and computational resources required increase
Solution Approach 1:
The patent segments the opinion data processing system into distinct frame extraction rules that target specific patterns and relationships. Each frame rule processes a particular aspect of opinion data (e.g., preferences, comparisons, attributes), dividing the complex task into manageable, specialized components that can be independently applied and maintained.
Solution Approach 2:
The patent introduces frame instances as intermediary structures that mediate between raw opinion data and graphical representations. These frame instances serve as structured intermediaries that capture extracted information in a standardized format, facilitating the transformation from unstructured text to visual graphs without requiring direct complex processing between raw data and visualization components.
2Measurement precision
If frame-based analysis is used to extract structured information from natural language, then the precision of preference measurement is improved, but the processing time and computational effort increase
Solution Approach 1:
The patent applies preliminary action by pre-defining frame extraction rules that encode domain knowledge about preference expressions, product attributes, and comparison patterns. These rules are prepared in advance and can be directly applied to opinion data without requiring complex real-time analysis, thus reducing processing time while maintaining measurement precision through structured extraction.
3Ease of operation
If graphical representations like Instance Graphs and Instance Plots are generated to visualize preferences, then the ease of understanding market dynamics is improved, but the device complexity and visualization computation increase
Solution Approach 1:
The patent extracts only the essential preference relationships and product attribute information needed for visualization from the full opinion data corpus. By taking out and focusing on specific relevant elements (preferences, comparisons, key attributes) rather than processing all data, the system generates clear graphical representations with reduced computational complexity while maintaining ease of understanding for end users.
Data Source
AI summary
The following graphical representations, of frame instance data, are presented: Instance Graph and Instance Plot. An Instance Graph is a kind of directed graph that represents directed relationships between items, as established by frame instances. An example frame is the Preference Frame, as applied to online opinion data. The degree or “influence” of a node can be graphically indicated. Multiple edges, between two nodes, can be represented as a compound edge. Each node can be modeled as having a field, causing it to repel all other nodes, which each edge can be modeled as producing an attractive force. The “net preference” of a node is the difference between its outdegree and indegree. From the “influence” and “net preference” values, for nodes of an Instance Graph, an Instance Plot can be produced. One axis of an instance plot is based on “influence” and another axis is based on “net preference.”