Process Input Relevance Identification via Stochastic Gradient Boosting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately identify the relevance of process inputs in industrial systems, often incorporating both signal and random noise, which can lead to exaggerated estimates of input influence and meaningless dependencies.
Innovation Solution
The method computes an input relevance measure by obtaining sensor measurements from a sensor array, partitioning the data, building stochastic gradient boosting models for each partition, and generating partial dependency plots for each process input. This approach estimates the degree of change in the process output due to varying the process inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods are used to identify process input relevance, then all sensor measurements are considered, but the measurement precision deteriorates due to inclusion of random noise and signal contamination
Solution Approach 1:
The data is segmented into multiple partitions, and multiple stochastic gradient boosting models are built for different partitions. This segmentation allows the system to process and evaluate sensor measurements in divided groups, reducing the impact of noise in any single partition and improving the overall precision of input relevance identification.
Solution Approach 2:
The method extracts and removes random noise and contaminated signals from the sensor measurements through the statistical aggregation process. By computing input relevance measures across multiple data partitions and models, the system effectively separates meaningful signal patterns from random noise, retaining only the relevant information.
2Reliability
If multiple data partitions and stochastic gradient boosting models are built for each partition, then the reliability of input relevance identification is improved, but the device complexity increases
Solution Approach 1:
The computational workload is segmented into multiple independent data partitions, each processed by its own stochastic gradient boosting model. This segmentation enables parallel processing and distributed computation, which improves reliability through ensemble aggregation while managing complexity by dividing the overall computational task into smaller, more manageable units.
Solution Approach 2:
Multiple stochastic gradient boosting models built on different data partitions are merged through aggregation to produce a final input relevance measure. This merging process combines the results from individual models, improving reliability through ensemble methods while the modular structure keeps each individual model's complexity manageable.
Data Source
AI summary
Computing an input relevance measure of process inputs by obtaining from a sensor array sensor measurements that relate process inputs to a process output, computing from the sensor measurements, a number of data partitions, building for each data partition of the number of data partitions a corresponding stochastic gradient boosting model, and computing for each process input, a number of partial dependency plots with each partial dependency plot being based on the corresponding stochastic gradient boosting model. The input relevance measure is then computed for each process input based on the number of partial dependency plots of the process input to estimate a degree of change in the process output obtained by varying the process input.