Spectroscopy and Chemometric Analyses

EXPERIMENTAL DESIGN ISSUES ASSOCIATED WITH CLASSIFICATIONS OF HYPERSPECTRAL IMAGING DATA

The paper identifies and addresses two experimental design issues in hyperspectral imaging classification: the 'small N, large P' problem and lack of independent validation data, proposing a method to improve model reliability.

Averaging a local PLSR pipeline to predict chemical compositions and nutritive values of forages and feed from spectral near infrared data

The paper extends PLSR averaging to a kNN-LWPLSR pipeline for predicting chemical compositions of forages and feed from NIR data, finding averaging improves performance.

Using transfer learning and dimensionality reduction techniques to improve generalisability of machine-learning predictions of mosquito ages from mid-infrared spectra

This study improves mosquito age prediction from mid-infrared spectra using transfer learning and dimensionality reduction, achieving high accuracy across different populations.

Classification of plant growth-promoting bacteria inoculation status and prediction of growth-related traits in tropical maize using hyperspectral image and genomic data

This study uses hyperspectral and genomic data to classify plant growth-promoting bacteria inoculation status and predict growth traits in maize, finding that hyperspectral data better predict shoot dry mass and stalk diameter while genomic data better predict plant height.

Near-infrared spectroscopy outperforms genomic selection for predicting sugarcane feedstock quality traits

This study compares genomic selection and near-infrared spectroscopy for predicting sugarcane feedstock quality traits, finding NIR outperforms genomics and their integration does not improve accuracy.

Validation of binary non-targeted approaches: mathematical framework and experimental designs

This paper develops a mathematical framework and experimental designs for validating binary non-targeted methods to improve their reliability and performance in food fraud detection.

Making WAVES in Breedbase: An Integrated Spectral Data Storage and Analysis Pipeline for Plant Breeding Programs

The paper develops an open-source R package, waves, for spectral data analysis in plant breeding and integrates it into the Breedbase database to enhance spectral data handling and analysis.

An Autoencoder and Artificial Neural Network-based Method to Estimate Parity Status of Wild Mosquitoes from Near-infrared Spectra

This paper develops an autoencoder and ANN model to estimate parity status of wild mosquitoes using NIR spectra, achieving high accuracy across multiple datasets.

AI-based identification of grain cultivars via non-target mass spectrometry

The paper uses AI and non-target mass spectrometry to distinguish wheat and spelt cultivars and validates the method's discriminatory power.

Functional data analysis techniques to improve the generalizability of near-infrared spectral data for monitoring mosquito populations

This paper develops a functional data analysis method to improve the generalizability of near-infrared spectral data for mosquito population monitoring across geographic regions.

Age Grading An. Gambiae and An. Arabiensis Using Near Infrared Spectra and Artificial Neural Networks

This study compares artificial neural networks and partial least squares regression for age grading malaria mosquitoes using NIRS, finding higher accuracy with direct binary classifiers.

Statistical loadings and latent significance simplify and improve interpretation of multivariate projection models

The paper introduces a strategy to improve OPLS model interpretation by using post-hoc linear regression and statistical t-values to enable multivariate analysis of variable significance.