User Routine Prediction via Location Data Cleaning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computing systems face challenges in accurately predicting a user's routine due to noise in location data from mobile devices, which is reported irregularly to conserve battery power, leading to a lack of effective models for routine prediction.
Innovation Solution
A computing system that processes location entries from mobile devices to identify visits, clean and anonymize data, and employ algorithms like frequent itemset mining, sequence-to-sequence models, or masked time-dilated convolutional networks to generate a predicted routine by clustering and labeling time intervals based on visited places, providing recommendations and notifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If mobile computing devices report location data continuously to predict user routines accurately, then prediction accuracy is improved, but battery power consumption increases
Solution Approach 1:
The system performs preliminary actions by collecting and storing location data during off-peak times or in advance, then processes this stored data to generate routine predictions. This allows the system to make accurate predictions without requiring continuous real-time location reporting, thereby reducing battery consumption while maintaining prediction accuracy.
Solution Approach 2:
The system maintains continuity of useful action by using historical location data and machine learning models to continuously improve prediction accuracy over time. Instead of requiring continuous location reporting, the system processes accumulated data periodically to refine predictions, maintaining prediction quality while reducing the frequency of location data collection.
2Measurement precision
If mobile computing devices report location data frequently to improve routine prediction accuracy, then prediction accuracy is improved, but data transmission overhead increases
Solution Approach 1:
The system extracts only the essential information from location data (such as key locations, timestamps, and movement patterns) rather than transmitting complete location trajectories. By extracting and transmitting only the critical data points needed for routine prediction, the system reduces data transmission volume while maintaining prediction accuracy.
Solution Approach 2:
The system performs preliminary processing of location data locally on the mobile device to identify and filter out redundant information before transmission. This preliminary action reduces the quantity of data that needs to be transmitted to the server, thereby reducing data transmission overhead while preserving the essential information needed for accurate routine prediction.
3Measurement precision
If the system processes and cleans location data to remove noise and anomalies, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The system performs preliminary data cleaning and noise removal operations during off-peak times or in advance, preparing the data structure before it is needed for prediction. By pre-processing the data to remove obvious noise and anomalies, the system reduces the processing time required during actual prediction operations while maintaining high prediction accuracy.
Solution Approach 2:
The system applies different processing quality levels to different parts of the data. Instead of uniformly processing all location data with high intensity, the system applies selective processing - using lighter processing for obvious noise removal and more intensive processing only for critical data segments, thereby reducing overall processing time while maintaining prediction accuracy.
Data Source
AI summary
Described herein are technologies related to generating a predicted routine of a user of a mobile computing device. Location entries generated by the mobile computing device are processed to generate visit entries, wherein the visit entries are representative of visits made by the user to places over several days. An input sequence of states is constructed based upon the visit entries, wherein each state has a place identifier assigned thereto, and further wherein each state corresponds to a time interval of predefined length. A predicted routine of the user is generated based upon the input sequence of states.