Query Classification Model for Search Relevance Optimization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Online retail stores face challenges in optimizing their search features to efficiently direct consumers to relevant products, leading to potential loss of business due to irrelevant search results, which can be improved by enhancing query classification models to generate category mappings.

Innovation Solution

A method that identifies and applies category mappings for received queries based on user interactions, using a query classification model to rank categories and assign importance levels, thereby improving search relevance by linking queries to relevant products and categories.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If human input is used to modify product fields to improve search results, then search relevance is improved, but significant effort and error-proneness increase

Engineering Contradiction:

Improvesearch relevanceVSAvoidhuman effort required

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system automatically modifies product fields and generates search optimizations without human intervention. The computer system performs self-service by autonomously analyzing query logs, identifying dominant queries, generating category mappings, and updating product fields based on learned patterns from user search behavior

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual human operations with an automated computational system. Instead of human operators manually modifying product fields, the system uses algorithms to process query logs, classify dominant queries, and automatically update product data structures, substituting mechanical human labor with automated software processes

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If more products are made available for purchase, then product variety increases, but difficulty in finding specific products increases

Engineering Contradiction:

Improveproduct varietyVSAvoidproduct location difficulty

Core Design Contradiction:

Adaptability or versatilityVSDifficulty of detecting and measuring

Solution Approach 1:

The system uses feedback from query logs and user search behavior to continuously improve product categorization and search mappings. By analyzing which queries lead to successful product finds and which don't, the system refines its category mappings and dominant query classifications, creating a self-improving feedback loop that adapts to user needs

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary classification of queries and pre-establishes category mappings before users need them. By proactively analyzing query patterns and pre-organizing products according to dominant queries and user behavior, the system prepares the search infrastructure in advance, making product location faster when users actually search

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10366092B2Search relevance

Publication Date: 2019.07.30 WALMART APOLLO LLC

AI summary

The present invention extends to methods, systems, and computer program products for training a classification model to predict categories. In one implementation, a method identifies category mappings generated for dominant queries associated with a query log. The method identifies mappings between a first set of queries and categories shown for the first set of queries, and identifies mappings between a second set of queries and clicked products for the second set of queries. A classification model is trained based on the mappings generated for dominant queries, the mappings between queries and the shown categories, and the mappings between queries and the clicked products.