August 18, 2025DiagnosticsOpen Access

Explainable AI-Based Feature Selection Approaches for Raman Spectroscopy

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Exploratory methods optimize feature selection and enhance accuracy in Raman spectroscopy models, indicating diverse application needs.

The proposed methods achieved comparable accuracy levels while using only 10% of the features, indicating effective reduction.
Using convolutional neural networks with GradCam provided the highest average accuracy in feature extraction.
Analysis included three real-world datasets across four classifiers, demonstrating robust performance of the methods.
No single approach outperformed in all cases, highlighting the necessity for tailored feature selection strategies.

Rossberg et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

August 18, 2025DiagnosticsOpen Access

View Full Paper

Ask AI

Bookmark

Authors

Exploratory methods optimize feature selection and enhance accuracy in Raman spectroscopy models, indicating diverse application needs.

The proposed methods achieved comparable accuracy levels while using only 10% of the features, indicating effective reduction.
Using convolutional neural networks with GradCam provided the highest average accuracy in feature extraction.
Analysis included three real-world datasets across four classifiers, demonstrating robust performance of the methods.
No single approach outperformed in all cases, highlighting the necessity for tailored feature selection strategies.

Rossberg et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark