Terminal Recommendation Architecture Using a Shared Algorithm Library
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile terminals face code redundancy issues due to the inclusion of machine learning algorithms in the service code of each intelligent service, leading to inefficiencies.
Innovation Solution
A method and terminal implementation that utilize a local algorithm library, allowing the terminal to invoke algorithms based on configuration parameters and user data, reducing the need for algorithm duplication in individual service codes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine learning algorithms are carried in service code of each intelligent service, then intelligent recommendation services can be implemented on the mobile terminal, but code redundancy is caused
Solution Approach 1:
The patent merges the machine learning algorithms from individual service codes into a shared algorithm library. The terminal obtains algorithms from this common library rather than embedding them in each service's code, thereby eliminating redundancy while preserving intelligent recommendation capabilities across multiple services.
Solution Approach 2:
The algorithm library serves as a universal resource that multiple intelligent services can access and utilize. By making the algorithm library multi-functional and shared across different services, the system avoids duplicating the same algorithms in each service code while maintaining adaptability and versatility of the recommendation system.
2Reliability
If machine learning algorithms are embedded in each service code, then each service can independently perform user preference prediction, but computational efficiency and resource utilization deteriorate
Solution Approach 1:
The patent combines the computational resources and algorithms into a shared library that multiple services access collectively. This merging approach maintains the ability to perform user preference prediction while improving computational efficiency by avoiding redundant algorithm execution and reducing overall resource consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of the present invention provide an intelligent recommendation method and a terminal. The method includes: receiving, by a terminal, a service request of a recommendation application, where the service request includes a configuration file, the configuration file includes at least an algorithm parameter and portfolio information that is used to identify an algorithm portfolio structure, and the recommendation application is included in the terminal; invoking, by the terminal, at least one algorithm from an algorithm library of the terminal based on the algorithm parameter and the portfolio information; predicting, by the terminal based on the at least one algorithm and user data of the terminal, at least one service currently required by a user of the terminal, where the user data of the terminal is stored in a personal database that is located in the terminal; and displaying, by the terminal, at least one icon corresponding to the at least one service. Code redundancy can be reduced in the embodiments of the present invention.