Service Resource Indicator Prediction With Time-Frequency Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing time sequence prediction methods capture a limited amount of information and are low in accuracy, affecting the applicability of service execution.
Innovation Solution
A method that combines time domain and frequency domain features to predict service resource indicators, using a time domain prediction model and a frequency domain prediction model to generate and weight prediction sequences, enhancing local dependency and global correlation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing prediction methods are used, then the prediction process is simple, but the prediction accuracy is low
Solution Approach 1:
The patent segments the prediction task into two distinct domains: time domain prediction and frequency domain prediction. By dividing the original complex prediction problem into separate time-based and frequency-based sub-tasks, each handled by dedicated prediction models, the system captures different aspects of the indicator sequence characteristics independently, thereby improving overall prediction accuracy while managing complexity through modular architecture
Solution Approach 2:
The patent introduces frequency domain analysis as an additional dimension beyond the traditional time domain approach. By transforming the indicator sequence into frequency domain features and applying separate prediction models in this new dimension, the system captures periodic patterns and spectral characteristics that are not visible in the time domain alone, thus enhancing prediction accuracy through multi-dimensional feature extraction
2Measurement precision
If more information is captured in prediction, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the comprehensive information extraction process into time domain feature extraction and frequency domain feature extraction performed by separate models. This segmentation allows each model to focus on specific types of information patterns, capturing diverse features efficiently without requiring a single overly complex model that would consume excessive computational resources
Solution Approach 2:
The patent merges the results from time domain prediction and frequency domain prediction through a weighted combination mechanism. By integrating the prediction sequences from both domains with appropriate weighting, the system achieves comprehensive information capture that leverages the strengths of both approaches while maintaining computational efficiency through the modular structure
Data Source
AI summary
A time sequence prediction method and apparatus for a service resource indicator, and a device. The method includes: obtaining a first indicator sequence monitored in a service, the first indicator sequence being used to characterize a measurement value of a specified resource indicator of the service within a historical preset time period; invoking a time domain prediction model to perform prediction based on a time domain feature of the first indicator sequence to obtain a first prediction sequence of the specified resource indicator within a future preset time period; invoking a frequency domain prediction model to perform prediction based on a frequency domain feature of the first indicator sequence to obtain a second prediction sequence of the specified resource indicator within the future preset time period; and weighting the first prediction sequence and the second prediction sequence to obtain a prediction sequence corresponding to the service.