Classifier-Based Synonym Discovery for Search Query Expansion

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Users face incomplete search results due to unawareness of domain-specific synonyms, as search interfaces require manual and costly processes to discover related words, which are specialized across various technical domains.

Innovation Solution

A related feature determination system using machine learning to automatically identify synonyms by training a classifier with a modified positive and negative set, allowing for the expansion of search queries with context-sensitive and recursive keyword determination.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If manual processes are used to discover and compose related words in a database, then search query completeness can be improved, but the process becomes expensive and time-consuming

Engineering Contradiction:

Improvesearch query completenessVSAvoidmanual discovery process time

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The system enables automatic discovery of related words by having the software itself perform the discovery process through machine learning classification, eliminating the need for human editors to manually compose the database of related words. The classifier automatically identifies synonyms and related terms from the document database.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The manual mechanical process of human editors discovering and composing related words is replaced with an automated machine learning system that uses classifiers to identify synonyms and related terms, substituting human cognitive work with computational processes.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If a database of related words composed by human editors is used, then search functionality can be improved, but the process becomes expensive and requires extensive human intervention

Engineering Contradiction:

Improvesearch functionalityVSAvoidmanual composition process

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The system performs self-service by automatically building and maintaining the database of related words through machine learning, without requiring human editors to manually compose or update the database. The classifier continuously learns from the document database to identify new synonyms and related terms.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system changes the fundamental parameter of how related words are discovered - shifting from human editorial judgment to automated machine learning classification. This parameter change transforms the process from manual composition to automatic identification through trained classifiers.

Inventive Principle:

Principle #35Parameter changes

3Loss of information

If users manually expand query terms with known synonyms, then search completeness improves, but the difficulty for users to recognize or learn domain-specific terms increases

Engineering Contradiction:

Improvesearch completenessVSAvoiduser awareness of domain terms

Core Design Contradiction:

Loss of informationVSDifficulty of detecting and measuring

Solution Approach 1:

The system provides feedback to users by automatically suggesting related words and synonyms that are relevant to their search queries. The classifier analyzes the document database to identify terms that users should be aware of, and presents these suggestions to improve their search queries without requiring users to learn domain terminology manually.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The machine learning classifier acts as an intermediary between the user's initial query and the document database, automatically identifying and suggesting related terms that bridge the gap between user knowledge and domain-specific terminology used in the documents.

Inventive Principle:

Principle #24Intermediary (Mediator)

4Adaptability or versatility

If conventional search interfaces use manually composed databases of related words, then synonym expansion works for known terms, but the process cannot keep up with specialized terminology across different technical domains

Engineering Contradiction:

Improvesynonym expansion capabilityVSAvoiddiscovery process efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system transforms the static, manually composed database of related words into a dynamic system that automatically adapts to new domains and terminology. The machine learning classifier continuously learns from the document database, enabling the system to adapt to specialized terminology across different technical domains without manual intervention.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The machine learning classifier provides a universal solution that works across multiple technical domains and specialized terminologies, replacing the need for separate manual composition processes for each domain. The system handles diverse domains uniformly through automated learning.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS8165972B1Determining a feature related to an indication of a concept using a classifier

Publication Date: 2012.04.24 MICRO FOCUS LLC

AI summary

A related feature determination system determines a feature related to an indication of a concept. A set of positive cases and a set of negative case are selected using the indication of a concept. A classifier is induced for the concept using the set of positive cases in a manner blinded to the indication of a concept. The set of negative cases is applied to the classifier. A feature related to the indication of a concept is determined using results of applying the classifier to the set of negative cases.