Short Message Classification System for Targeted Content Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mechanisms for managing short messages, such as tweets, lack effectiveness in facilitating more advanced interactions and content distribution beyond subscribed users, limiting their utility for businesses and other promoting entities.
Innovation Solution
The development of methods and apparatus for analyzing and classifying short messages using Gradient Boosted Decision Trees and Latent Dirichlet Allocation algorithms to identify message classes, allowing for the retention and distribution of classified content to non-subscribers based on user interests and online activities, enabling targeted promotion across various platforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If short messages are sent only to subscribing users, then message delivery reliability is improved, but message reach and promotional effectiveness deteriorate
Solution Approach 1:
The patent segments the message distribution system into two distinct channels: subscribed users who receive all messages reliably, and non-subscribers who receive selectively distributed messages based on classification. This segmentation allows the system to maintain high delivery reliability for subscribers while expanding reach to non-subscribers through targeted content distribution.
Solution Approach 2:
The patent introduces an intermediary classification system that acts as a mediator between the message source and potential recipients. The classification apparatus analyzes message content and determines appropriate distribution targets, enabling reliable delivery to subscribers while selectively extending reach to non-subscribers based on message relevance and user interests.
2Productivity
If message classification and analysis systems are implemented, then content distribution effectiveness is improved, but system complexity deteriorates
Solution Approach 1:
The classification system is segmented into distinct functional modules: message reception component, classification component using multiple models, and distribution component. Each module performs a specific function, making the overall complex system manageable and maintainable while achieving high content distribution effectiveness through specialized processing at each stage.
Solution Approach 2:
The classification system employs multiple classification models that can be applied universally across different message types and promoting entity accounts. These models use lexicons and training sets that can be adapted to various domains, allowing the system to handle diverse content effectively without requiring entirely separate systems for each use case.
3Measurement precision
If multiple classification models are used for message analysis, then classification accuracy is improved, but processing time deteriorates
Solution Approach 1:
The system performs preliminary actions by pre-processing messages and preparing classification data before actual classification occurs. Training sets and lexicons are prepared in advance, and the multiple classification models are pre-configured, allowing rapid sequential evaluation of messages against multiple criteria without significant processing delays during live operation.
Solution Approach 2:
The patent merges multiple classification models into a unified classification framework that processes messages through various models in sequence or parallel. The results from multiple models are combined to produce a final classification decision, achieving high accuracy through ensemble reasoning while managing processing time through efficient model integration and result aggregation.
Data Source
AI summary
Disclosed are methods and apparatus for analyzing and using online short messages from promoting entity accounts (e.g., business or non-profit accounts). In one embodiment, a method of analyzing and using messages sent for a plurality of promoting entity accounts is disclosed. A plurality of models for classifying a plurality of messages based on a plurality of message features are obtained for each message. Each message is sent via a computer network between a selected one of the promoting entity accounts and one or more subscribing users that subscribe to receive messages from such selected promoting entity account, and each model is trained to identify whether a message belongs to a particular class based on a lexicon that was generated for such particular class and a training set of messages that belong to the particular class and message that do not belong to the particular class. A new message is classified based on the models and retaining classification information regarding the new message in a database that is accessible by a user so as to review the classification information on a computer display.