Spectroscopic Model Updating With Master-Target Cross-Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Spectroscopic models become inaccurate over time due to changes in raw materials and variations in spectrometer calibrations and environments, leading to inaccuracies in raw material identification and quantification.
Innovation Solution
A cross-validation technique is employed to merge data from a master data set with a target data set for training and validation, optimizing the partial least squares factor to generate an updated spectroscopic model, improving accuracy and reducing the need for individual spectrometer-specific master data sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a spectroscopic model is updated using only target data set, then the model adapts to new conditions, but the model accuracy deteriorates due to limited data representation
Solution Approach 1:
The patent combines the master data set (comprehensive historical data) with the target data set (new condition data) to create an updated spectroscopic model. This merging ensures the model maintains accuracy by leveraging the robustness of historical data while adapting to new conditions through target data, resolving the contradiction between adaptability and accuracy.
Solution Approach 2:
The updated spectroscopic model serves multiple functions: it maintains compatibility with historical data patterns while simultaneously adapting to new measurement conditions. The model is designed to handle both legacy and current data characteristics, making it universally applicable across different time periods and conditions.
2Measurement precision
If multiple spectrometer-specific master data sets are collected, then model precision is maintained for each device, but deployment costs increase
Solution Approach 1:
The patent creates a single updated spectroscopic model that can be universally deployed across multiple spectrometers without requiring individual spectrometer-specific master data sets. This universal model maintains precision across different devices by incorporating diverse data from the master data set, thereby reducing deployment complexity and costs.
Solution Approach 2:
The approach segments the data collection requirement by using a shared master data set that represents multiple spectrometers' characteristics, rather than requiring complete data sets for each individual device. This segmentation reduces the overall data collection burden while maintaining individual device precision.
3Measurement precision
If cross-validation is performed with merged data sets, then model calibration accuracy improves, but computational time increases
Solution Approach 1:
The patent performs preliminary cross-validation analysis to determine the optimal number of principal components before final model training. This preliminary action identifies the most informative data subsets and validation strategies, reducing the computational burden of subsequent full model training while maintaining calibration accuracy.
Data Source
AI summary
A device may receive a master data set for a first spectroscopic model; receive a target data set for a target population associated with the first spectroscopic model to update the first spectroscopic model; generate a training data set that includes the master data set and first data from the target data set; generate a validation data set that includes second data from the target data set and not the master data set; generate, using cross-validation and using the training data set and the validation data set, a second spectroscopic model that is an update of the first spectroscopic model; and provide the second spectroscopic model.