Classification Model for E-commerce Query Category Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online consumers face difficulties in navigating and finding specific products on e-commerce websites due to the vast number of available items, leading to inefficient search results and potential loss of business for retailers, as current search features often require significant human input and are error-prone.
Innovation Solution
A classification model is trained to predict categories by analyzing query logs, identifying mappings between queries and categories, and between queries and clicked products, using a computer architecture that includes a query classification module to calculate selection rates, category scores, and rank categories, thereby optimizing search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human input is used to modify product fields to improve search relevance, then search result accuracy is improved, but labor cost and time consumption increase significantly
Solution Approach 1:
The system enables self-service by automatically generating category mappings through machine learning models that analyze query logs and product data without human intervention. The classification model autonomously modifies product fields and generates category assignments, eliminating the need for manual human input while maintaining high search result accuracy.
Solution Approach 2:
Manual human operations are replaced with automated machine learning systems. The patent substitutes human analysts who manually modify product fields with computational models that automatically analyze query logs, product attributes, and category structures to generate accurate category mappings, thereby eliminating time-consuming manual labor.
2Measurement precision
If human input is used to modify product fields to improve search relevance, then search result accuracy is improved, but error rate increases due to human error
Solution Approach 1:
The system performs self-service through automated machine learning models that consistently apply classification rules without human error. The models analyze query logs and product data systematically, eliminating the variability and mistakes inherent in manual human operations while maintaining high accuracy through algorithmic consistency.
Solution Approach 2:
The patent replaces human manual operations with automated computational systems. Machine learning models process query logs and product attributes through consistent algorithmic logic, eliminating human errors such as typos, fatigue-related mistakes, and inconsistent judgment, thereby improving reliability while maintaining accuracy.
3Productivity
If automated search algorithms are used without human input, then processing speed is improved, but search result relevance deteriorates
Solution Approach 1:
The system performs preliminary action by pre-training classification models on extensive query logs and product data before actual search operations. The models are pre-trained to understand category structures and product attributes, enabling them to quickly generate accurate category mappings during actual search operations without requiring real-time human intervention, thus achieving both speed and relevance.
Solution Approach 2:
The system implements feedback mechanisms where classification models continuously learn from query logs and search performance data. The models analyze user behavior patterns, click-through rates, and search results to refine their category mapping accuracy over time, enabling automated systems to improve relevance while maintaining high processing speed through iterative learning.
Data Source
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.