Geo-Block Location Prediction Using Feature Space Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current location-based information technologies face challenges in accurately predicting mobile device locations in real-time, particularly in translating raw location data into meaningful signals for providing relevant information to mobile users, due to limitations in existing systems for processing and filtering location events across time and space.
Innovation Solution
A location prediction system that utilizes geo-blocks and geo-fences to process location events, constructing feature spaces for machine learning models to predict mobile device locations, and dynamically adjusts targeting areas based on relevance measures and pacing status of information campaigns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional location-based information technologies process raw location data directly, then comprehensive location information is captured, but processing time increases and prediction accuracy decreases
Solution Approach 1:
The patent segments raw location data into structured location events with specific attributes (timestamp, geo-fence ID, geo-block ID, direction, speed). This segmentation organizes continuous location streams into discrete, meaningful units that can be processed efficiently by machine learning models, improving both prediction accuracy and processing speed.
Solution Approach 2:
The system performs preliminary processing by pre-defining geo-fences and geo-blocks, and pre-structuring location events before they are fed into prediction models. This preliminary organization of spatial data and event formats reduces the computational burden during real-time prediction, addressing the time-loss issue while maintaining accuracy.
2Measurement precision
If detailed location events are processed and stored for machine learning, then prediction accuracy improves, but memory and processing requirements increase
Solution Approach 1:
The patent extracts only the most relevant features from raw location data, such as geo-fence entry/exit events, geo-block transitions, directional information, and speed metrics. By taking out and focusing on these key predictive features rather than processing all raw location coordinates, the system maintains high prediction accuracy while reducing memory and computational requirements.
Solution Approach 2:
The system transforms raw location parameters (coordinates, timestamps) into standardized event parameters (geo-fence ID, geo-block ID, direction, speed). This parameter transformation converts continuous, high-volume location data into discrete, structured events that are more efficient for machine learning processing, balancing accuracy with reduced complexity.
3Productivity
If location data is filtered and structured into events, then processing efficiency improves, but data completeness may be reduced
Solution Approach 1:
The patent creates a universal location event structure that can represent multiple types of location information (geo-fence events, geo-block events, directional changes, speed variations) within a single standardized framework. This multi-functional event structure ensures that diverse location data is captured efficiently without loss of information, while maintaining processing efficiency through uniform handling.
Data Source
AI summary
A system coupled to a packet-based network is configured to predict the locations of mobile devices that have communicated with the packet-based network. The system includes a request processor configured to detect location events associated with mobile devices communicating with the packet-based network, each location event corresponding to a time stamp and identifying a geo-place in a geo database. The geo-places include geo-blocks and geo-fences. The system further comprises a location prediction subsystem configured to construct first feature space using first location events and second feature space using second location events, and to extract a set of labels from third location events. The location prediction subsystem is further configured to train a prediction model using the first feature space and the set of labels, and to apply the prediction model to the second feature space to obtain prediction results.