Rank-Based Keyword Search Using User Behavior Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing keyword search methods in documents are inefficient as they do not account for user preferences and behaviors, leading to a need for more accurate and efficient search results that direct users to specific sections of interest within documents.
Innovation Solution
A rank-based keyword search method that collects and utilizes user information such as object search data, holding time, and selecting time to calculate a priority-based search, allowing users to view preferred keywords first, and dynamically updates based on user behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional keyword search methods are used to find all instances of a word or phrase in a document, then completeness of search results is improved, but search time and user effort increase significantly
Solution Approach 1:
The patent applies local quality by differentiating the importance of different keyword instances based on their contextual location within the document. Instead of treating all instances equally, the system assigns different weights to keywords appearing in different sections (e.g., abstract, introduction, conclusion, body paragraphs), allowing the search to focus on high-value regions while maintaining completeness when needed.
Solution Approach 2:
The patent changes the parameter of keyword importance from a uniform value to a variable parameter that depends on contextual factors such as section location, frequency of occurrence, and proximity to other relevant terms. This allows the search system to dynamically adjust which instances are prioritized, reducing the effective search space while preserving the ability to find all instances if required.
2Ease of manufacture
If keyword search returns all matched instances in sequence of appearance, then simplicity of implementation is improved, but relevance to user needs deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing metadata about keyword instances before the actual search occurs. This includes pre-computing section identifiers, frequency counts, and contextual weights for each keyword instance. When a search is performed, this pre-processed information allows for immediate ranking and filtering without complex real-time analysis, maintaining implementation simplicity while improving relevance.
3Stability of the object's composition
If uniform treatment of all keyword instances is used, then consistency of search behavior is improved, but adaptability to different user preferences deteriorates
Solution Approach 1:
The patent applies dynamics by making the search results adaptive to user preferences while maintaining a consistent underlying framework. The system allows users to configure their preferences (e.g., prioritize abstracts, prefer certain sections, adjust weighting schemes) and dynamically adjusts the ranking of keyword instances based on these preferences. The core search algorithm remains consistent, but the presentation and prioritization of results are flexible and adaptive.
Data Source
AI summary
Embodiments of the present invention provide efficient systems and methods for searching a keyword. Embodiments of the present invention can be used to more quickly and comprehensively search a document for a keyword or phrase, based on the preferences and behaviors of a user, so that the returned results to the user are ranked in order of the preferences of the user. By factoring in user data such as time spent on certain content and highlighted content, the system is able to more efficiently return preferred results of a user of a keyword search.