Automatic Search Dictionary Update via Vectorization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Maintaining up-to-date search dictionaries for websites is a manual and labor-intensive task, requiring website owners to manually identify and link search keywords that result in null or sub-optimal search results.

Innovation Solution

A digital data processing system that includes a server and client devices, utilizing a search engine and dictionary curation software to automatically or semi-automatically update site-specific dictionaries by analyzing search logs and incorporating related terms from a corpus lookup table, with the help of vectorization tools like GloVe and Word2Vec to determine term relatedness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual updating of search dictionaries is performed, then search result accuracy can be maintained, but labor intensity and time consumption increase significantly

Engineering Contradiction:

Improvesearch result accuracyVSAvoidtime consumption

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automatic self-updating of search dictionaries by analyzing search logs and using vectorization algorithms to identify related terms, eliminating the need for manual intervention while maintaining accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Manual mechanical processes of dictionary updating are replaced with automated computational processes including log analysis, vectorization (GloVe, Word2Vec), and algorithmic term relationship detection

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If manual dictionary updates are performed, then search term alignment can be maintained, but operational complexity increases

Engineering Contradiction:

Improvesearch term alignmentVSAvoidoperational complexity

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The system automatically maintains search term alignment through self-service mechanisms that analyze search patterns and update dictionaries without requiring operator intervention

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system uses feedback from search logs to automatically detect null or sub-optimal search results and uses vectorization to identify appropriate term relationships, creating a closed-loop system that maintains alignment automatically

Inventive Principle:

Principle #23Feedback

3Productivity

If automated updating using vectorization tools is implemented, then productivity increases, but system complexity increases

Engineering Contradiction:

Improvedictionary update efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

Manual dictionary updating is replaced with automated computational systems using vectorization algorithms (GloVe, Word2Vec) that process search logs and generate term relationships algorithmically

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

Vectorization tools serve as intermediaries between search logs and the search dictionary, automatically translating search pattern data into structured term relationship information without direct human intervention

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10789293B2Automatic search dictionary and user interfaces

Publication Date: 2020.09.29 SALESFORCE INC

AI summary

A method of filtering content from a data set includes accepting a search request directed to a data set associated with a site, the search request including a search term that is not among terms represented in a site-specific lookup table representing site-specific relatedness of terms in that data set, such related terms including any of synonyms, hypernyms and hyponyms; generating an approximating lookup table by applying a transformation function to a corpus lookup table, the corpus lookup table representing relatedness, in a general corpus, of terms in the data set; identifying terms represented in the approximating lookup table that are related terms of the search term; and filtering from the data set digital content that includes any of the search term and the terms identified from the approximating table as related terms of the search term.