Machine Learning Search Session Analysis for Result Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional search engines struggle to provide relevant results as they do not effectively capture and analyze multiple navigations across different webpages during a search session, leading to inefficiencies and lower satisfaction for users, as they fail to understand the user's intent and adapt search results dynamically.
Innovation Solution
A machine learning algorithm is trained to identify search starting and terminating events, track intermediary events across multiple webpages, and derive dynamic scores to understand user intent, thereby providing refined search results that are more relevant and efficient by analyzing navigation routes and user characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional search engines use basic search algorithms to provide search results, then the system complexity is low, but the search result relevance and user satisfaction are insufficient
Solution Approach 1:
The patent replaces traditional mechanical search algorithms with a machine learning-based system that uses neural networks to analyze user behavior patterns, navigation routes, and session data. This substitution enables the system to achieve higher search result relevance by learning from historical data while managing complexity through automated modeling rather than manual rule-based approaches
Solution Approach 2:
The machine learning system performs self-training by automatically analyzing user search patterns, navigation behavior, and session data to improve its own performance. The system learns from historical search sessions and continuously refines its models without requiring manual reconfiguration, enabling it to adapt to changing user needs while maintaining operational simplicity
2Measurement precision
If search engines analyze multiple navigations across webpages to understand user intent, then the search result accuracy improves, but the data processing complexity increases
Solution Approach 1:
The patent segments the search session into distinct components including search starting events, intermediary events (navigations), and search terminating events. By dividing the complex user journey into manageable segments, the system can process and analyze each component separately using machine learning models, reducing overall processing complexity while maintaining comprehensive intent understanding
Solution Approach 2:
The system performs preliminary analysis of search sessions by pre-processing and storing navigation data, user behavior patterns, and session information in a structured format. This preliminary action enables the machine learning models to quickly query and match against historical data during actual search operations, reducing real-time processing complexity while maintaining high accuracy
3Productivity
If the search engine tracks the entire search session including multiple webpages, then the search-to-end result success rate improves, but the time required for analysis increases
Solution Approach 1:
The system performs preliminary tracking and storage of search session data during user interactions, organizing navigation routes and behavior patterns in advance. This preliminary action allows the machine learning models to quickly retrieve and analyze relevant historical patterns during actual search queries, reducing real-time analysis time while maintaining comprehensive session understanding
Solution Approach 2:
The patent implements feedback mechanisms where the system analyzes search session outcomes and uses this information to refine its machine learning models. By continuously learning from actual user behavior patterns and search success rates, the system optimizes its analysis algorithms to achieve higher success rates with reduced computational time through iterative improvement
Data Source
AI summary
Techniques for training and/or using a machine learning (ML) algorithm to generate search results are disclosed. An ML algorithm is configured (i) to identify a search starting event and a search terminating event for a search session, where the search session includes multiple navigations across multiple different webpages, and (ii) to derive a dynamic score for the search session, where the dynamic score reflects whether the search session successfully identified an end result that was initially unknown at a time when the search starting event occurred. The trained ML algorithm can then be used during later search sessions to promote better search results and help users identify end targets or results in a faster and more intuitive manner.