Mobile Terminal Activity Recognition Using Server Prior Probabilities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for recognizing and forecasting user activities on mobile terminals are inaccurate due to limited data samples, leading to a poor user experience, as they rely solely on terminal data without considering broader activity patterns.
Innovation Solution
A method that collects terminal information, including position, and obtains prior probability information from a server to improve activity recognition and forecasting, incorporating this data into classification and forecasting models using algorithms like linear regression and decision trees.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only terminal application data is collected for user activity recognition, then the system complexity is low, but the recognition accuracy is insufficient due to limited data samples
Solution Approach 1:
The patent combines terminal-collected application data with server-provided prior probability data to form a comprehensive data set for activity recognition. This merging of data sources increases the number of data samples and improves recognition accuracy without requiring the terminal alone to handle all data collection complexity
Solution Approach 2:
The server acts as an intermediary that provides prior probability information to the terminal. This intermediary supplies additional contextual data (prior probabilities of different activities at different times and locations) that enhances the terminal's recognition capability without requiring the terminal to independently collect and process all necessary data
2Loss of information
If only terminal data is used for activity forecast, then the information pushed to user is generic, but the relevance to user interests is low
Solution Approach 1:
The system uses prior probability information from the server as feedback to improve the accuracy of activity recognition and forecasting. This feedback mechanism allows the terminal to adjust its predictions based on statistically derived prior probabilities, making pushed information more relevant to actual user activities and interests
Solution Approach 2:
The patent changes the parameters used for activity recognition by incorporating prior probability values that represent the likelihood of different activities occurring at specific times and locations. This parameter enhancement allows the system to adapt to different contexts and provide more personalized, relevant information to users
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
Embodiments of the present invention disclose a method for processing user information, a mobile terminal, and a server. The method includes: collecting, by a mobile terminal, terminal information, where the terminal information includes a position of the mobile terminal; obtaining prior probability information of different types of user activities at current time and in the position; and recognizing or forecasting, according to the prior probability information and the terminal information, an activity of a terminal user using the mobile terminal. In the embodiments of the present invention, an activity of a terminal user is recognized or forecasted with reference to not only terminal data of the terminal user but also prior probability information of different types of user activities. During the recognition or the forecast, the number of data samples increases, and therefore accuracy of the recognition or the forecast is improved; in addition, a difference between information that is pushed to the terminal user by applying a recognition or forecast result in the embodiments of the present invention and information in which the user is interested is comparatively small, and therefore terminal using experience of the user can be improved.