Leveraging Data Science Tools and Algorithms to Unveil New Insights on Historical Research Data: A Case of Insect Morphometrics

Résumé

Abstract The move towards open access and re-use of scientific research data is rapidly being embraced by the research community as best practice. Many research institutions are adopting a set of global data policy guiding principles to make data Findable, Accessible, Interoperable and Reusable (FAIR). This study is product of good research data stewardship of open access and re-use. We explored the use and application of advanced data science with machine learning tools and algorithms on historical data of insect morphometrics that were previously analyzed using conventional statistical methods, principal component analysis and canonical variate analysis. Herein, we assess the predictive performance of four machine learning classifiers; K-nearest neighbor (KNN), random forest (RF), support vector machine (the linear, polynomial and radial kernel SVMs) and artificial neural networks (ANNs) on the historical data of fruit fly morphometrics. KNN and RF performed poorly with overall model accuracy lower than “no-information rate” (NIR) (p-value>0.1). The SVM models had a predictive accuracy of >95% and Kappa >0.78 with accuracy significantly higher than NIR, p<0,001; while ANN model had a predictive accuracy of 96% and Kappa of 0.83 with accuracy also greater than NIR. We conclude that SVM and ANN models could be used to discriminate fruit fly species based on wing vein and tibia length measurements or any other morphologically similar pest taxa. These algorithms could be used as candidates for developing an integrated and smart application software for insect discrimination and identification.

Citer ce document

Accès au document

Ce document est hébergé sur son dépôt institutionnel d'origine.

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0