Dynamic Electronic Message Prioritization Using Unsupervised Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic message management systems struggle to effectively prioritize messages without making changes to the internal architecture of electronic devices, particularly failing to handle brief or lengthy messages with sparse or noisy data.
Innovation Solution
A method and system that dynamically prioritizes electronic messages using language and text processing to identify prioritization parameters, assigning values based on a predetermined order, allowing for prioritization without altering the device's internal architecture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If supervised learning algorithms are used to prioritize emails based on classifier features, then email prioritization can be automated, but the system fails to work well for brief messages with sparse data or long messages with noisy data
Solution Approach 1:
The patent changes the parameters used for email prioritization from traditional supervised learning features to unsupervised learning features including word frequency, phrase frequency, sender information, recipient information, and email metadata. This parameter transformation allows the system to effectively process both brief and long messages without relying on labeled training data, thereby maintaining automation while improving measurement precision across diverse message types
Solution Approach 2:
The patent substitutes the supervised learning mechanism with an unsupervised learning mechanism. Instead of using a classifier that requires labeled training data, the system employs unsupervised algorithms that automatically discover patterns in email data without human intervention, replacing the mechanical classification process with a more adaptive pattern recognition approach that works across all message lengths
2Measurement precision
If users manually scan their inbox and check sender details to prioritize messages, then message importance can be identified, but users spend excessive time on message processing
Solution Approach 1:
The patent implements self-service by enabling the email system to automatically prioritize messages without user intervention. The unsupervised learning algorithm autonomously analyzes email content, sender information, recipient information, and metadata to determine priority, allowing the system to serve itself rather than requiring users to manually evaluate each message's importance
Solution Approach 2:
The patent applies preliminary action by pre-calculating and storing priority scores for emails based on unsupervised learning analysis. When emails arrive or are stored, the system proactively computes their importance using features like word frequency and sender information, so that when users view their inbox, messages are already prioritized and ready for quick identification without requiring users to perform the analysis themselves
Data Source
AI summary
The present disclosure relates to a method for dynamically prioritizing electronic messages in an electronic device. In an embodiment an electronic message prioritization unit configured in an electronic device receives one or more electronic messages from other electronic device. The prioritization unit analyzes each of the received electronic messages using language and text processing logic to identify prioritization parameters. The prioritization unit assigns a value in a predetermined order to each of the prioritization parameters based on a predefined prioritization order of the prioritization parameters. The electronic message prioritization unit prioritizes the electronic messages based on the assigned values.