Snake species identification by using natural language processing

Abstract

The paper presents the snake species identification by using natural language processing. It aims to help medical professionals in predicting the snake species for snake-bite treatments based on the patient’s description of the snake. The decision in suitable anti-venom critically depends on the type of snake species. Wrong anti-venom may result in severe morbidity and mortality. This research investigates the human perception and the selection of words in describing a snake based on their visual view. The descriptions were presented in unstructured text, and the NLP processing involves pre-processing, feature extraction and classification. Four machine learning algorithms (naïve Bayes, k-Nearest Neighbour, Support Vector Machine, and Decision Trees J48) were used during training and classification. Our results show that J48 algorithm obtained the highest classification accuracy of 71.6% correct prediction for the NLP-Snake data set with high precision and recall.