Distributed Rule Base for Mobile User Intent Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy systems for collecting Internet user demographics and predicting user intentions based on spatial information fail to scale with the number of users and devices, leading to increased burdens on centralized computing infrastructures and limited capabilities in determining user activity and intentions across multiple locations.
Innovation Solution
The implementation of a system that delivers a rule base to a mobile device using location beacon identification values, allowing for real-time spatial information processing to predict user interests and intentions, reducing computer memory usage, processing power demands, and communication overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If centralized systems process spatial information for all users, then comprehensive user profiling is achieved, but system scalability deteriorates as user numbers increase
Solution Approach 1:
The patent divides the centralized processing system into distributed components where each mobile device executes local rule sets independently. The classification logic is segmented and deployed to edge devices (mobile phones), allowing parallel processing across thousands of devices simultaneously, thus maintaining profiling accuracy while enabling system scalability.
Solution Approach 2:
The patent transitions from a single centralized processing dimension to a multi-dimensional distributed architecture. Processing occurs across spatial dimensions (multiple locations), temporal dimensions (real-time updates), and hierarchical dimensions (device-level, network-level, and cloud-level processing), enabling both precision and scalability.
2Measurement precision
If more spatial and temporal data is collected from multiple locations, then user intent prediction accuracy improves, but computational resource requirements increase
Solution Approach 1:
The patent pre-compiles and distributes classification rule sets to mobile devices before they are needed for processing. By preparing the computational logic in advance and storing it locally, the system enables real-time intent prediction without requiring heavy computational resources during actual data processing, thus improving accuracy while conserving processing power.
Solution Approach 2:
The patent creates and distributes copies of classification rules to multiple mobile devices. Instead of one device processing all data centrally, each device receives a copy of the necessary rule sets and processes its own local data independently, reducing the computational burden on any single device while maintaining high prediction accuracy through distributed processing.
3Productivity
If rule sets are transmitted to mobile devices, then processing efficiency improves, but communication overhead increases
Solution Approach 1:
The patent transmits only the specific rule sets relevant to each mobile device's location, user profile, and context rather than sending complete rule bases to all devices. This localized approach ensures each device receives minimal necessary data for its specific processing needs, improving efficiency while minimizing communication overhead.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
A method, system, and computer program product for classifying user intent based on spatial information relayed via a mobile device. Embodiments commence upon receiving a location beacon identification value originating from a location beacon, which location beacon identification value is then relayed to the server from the mobile device. The server determines a set of one or more rules based at least in part on the location beacon identification value, and then transmits at least a portion of a set of one or more rules to the mobile device, (e.g., using a mobile device carriers infrastructure). The rules comprise triggers, and when a trigger fires, user categorization is determined and user intent is predicted. The user categorization and/or predicted user intent is used to select an advertising message.