Semantic Position Identification for Text Categorization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data management systems face inaccuracies in categorizing textual data into topics due to limitations in classifier-based techniques, which affect tasks such as data storage, search, and routing.

Innovation Solution

The system identifies semantic positions of text items in a continuous n-dimensional space, using a position-based feature dictionary to determine semantic positions of sentences and higher-level text portions, and assigns topics based on these positions through various algorithms like greedy, genetic, and clustering techniques.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If classifier-based techniques are used to assign topics to textual data, then the system can perform automated categorization, but the classification accuracy deteriorates

Engineering Contradiction:

Improveautomated categorizationVSAvoidclassification accuracy

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The patent transforms the traditional flat classification approach into a multi-dimensional semantic space where text portions are represented as vectors with multiple attributes. This dimensional transformation allows the system to capture nuanced semantic relationships that single-dimension classifiers miss, thereby maintaining automation while improving accuracy through richer feature representation.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The system dynamically adjusts classification parameters by computing semantic positions based on multiple features including word frequencies, phrase patterns, and contextual relationships. By changing from static classifier parameters to dynamic semantic parameters derived from actual text content, the system achieves both automation and high classification accuracy.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If simple classification algorithms are used, then the processing speed is high, but the ability to capture semantic similarity deteriorates

Engineering Contradiction:

Improveprocessing speedVSAvoidsemantic similarity capture

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent segments text into portions and represents each portion as a vector in semantic space, breaking down the complex task of semantic analysis into manageable computational units. This segmentation allows efficient processing of individual text portions while maintaining the ability to capture semantic relationships through vector operations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system replaces traditional mechanical classification algorithms with mathematical vector operations in semantic space. By substituting rule-based mechanical classification with continuous vector space mathematics, the system achieves both computational efficiency and superior semantic similarity capture through distance and angle calculations.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS8359190B2Identifying semantic positions of portions of a text

Publication Date: 2013.01.22 MICRO FOCUS LLC

AI summary

A computer executable method includes determining semantic positions in a semantic space of portions of a text. Once the semantic positions of portions of the text have been determined, an action is performed with respect to the text based on the determined semantic positions.