Abstract Habitat and morphometric information can be used as factors to differentiate species, primarily for species that are habitat Specific. Using the above concept a predictive model was created for the identification of tiger beetles using habitat and morphometric data. In this process different machine learning based classification algorithms (both single and ensemble) were evaluated to identify tiger beetle species based on their habitat and morphometric data. By considering each specimen collected as a record, a dataset of 468 records with 13 attributes (location, habitat and morphometric data) was created of 14 ground-dwelling tiger beetle species. Analysis of the results gained from different machine learning models revealed that the Extra Tree Classifier and Random Forest algorithms which are ensemble algorithms perform better than single classification models. Hence it’s proven that ensemble models have a positive effect on the overall quality of predictions, in terms of accuracy, generalizability ,lower misclassification costs and more stable than single classifiers. Ensemble Extra Tree Classifier and Random Forest algorithms have given all most the same overall accuracy (85%) with less than 0.12% difference. However, when consider both computational time with performance, ensemble Extra Trees Classifier can consider as the most suitable algorithm for the scenario. Although in most cases feature selection improves the classification accuracy, during the present scenario it became untrue. The main reason for above outcome can be, that so that optimum accuracy can only be gained by combining all the features.
Abeywardhana et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: