Dynamic Quality Metrics Forecasting via Neural Network Bias Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Organizations face challenges in monitoring and managing quality metrics across different phases of software development life cycles, as existing systems struggle to effectively track and predict changes in performance metrics due to varying influencing factors and complex relationships between them.
Innovation Solution
A computer-implemented method and system that receives parameter data from multiple sources, generates quality metrics, identifies relationships between these metrics, and predicts future values using a deep neural network with dynamic bias control, allowing for proactive quality management and optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional quality metric monitoring systems are used, then implementation is simple, but the system cannot effectively predict changes in performance metrics due to varying influencing factors
Solution Approach 1:
The system dynamically adjusts the monitoring model by continuously learning relationships between quality metrics and influencing factors. The neural network adapts to changing patterns in the data, allowing the system to evolve its prediction capabilities as new data becomes available, thereby improving reliability without requiring complete system replacement
Solution Approach 2:
The patent introduces a neural network as an intermediary layer between raw quality metric data and predictive insights. This intermediary learns and captures complex relationships between multiple influencing factors and quality metrics, enabling accurate predictions while keeping the overall system architecture manageable and interpretable
2Measurement precision
If comprehensive parameter data from multiple sources is collected, then prediction accuracy improves, but data processing complexity increases
Solution Approach 1:
The system merges multiple data sources and parameter types into a unified analysis framework. By combining defect data, process data, and quality metric data into a single neural network model, the system achieves comprehensive measurement precision while managing data processing complexity through integrated rather than separate handling of each data source
Solution Approach 2:
The patent transforms raw parameter data from multiple sources into meaningful features that the neural network can process efficiently. By changing the representation of input parameters into appropriate formats and feature sets, the system maintains high measurement precision while reducing the computational complexity of processing diverse data sources
3Adaptability or versatility
If dynamic relationships between quality metrics are identified, then proactive quality management is enabled, but computational requirements increase
Solution Approach 1:
The system performs preliminary learning of relationships between quality metrics during training phases, capturing dynamic patterns in advance. This preliminary action enables the model to make predictions during operation without requiring intensive real-time computation, thereby providing adaptability while managing computational resource consumption during actual quality management tasks
Data Source
AI summary
Systems and methods for dynamic quality metrics forecasting and management are provided. In embodiments, a method includes receiving, by a computing device, parameter data from one or more data sources for selected parameters, wherein the parameters are associated with one or more processes; generating, by the computing device, output values for plural quality metrics based on the parameter data; identifying, by the computing device, relationships between the plural quality metrics based on changes in the received parameter data and output values for the plural quality metrics over time; receiving, by the computing device, user-selected values for the selected parameters; and generating, by the computing device, predicted output values for the quality metrics based on the identified relationships between the quality metrics and the user-selected values.