A Data Mining Study for Detection of Venomous Snakes

Abstract

Snakes are one of the animals that people are most afraid of. Especially in rural areas, people and snakes share the same environment. Therefore, they had to develop methods of protection for these dangerous animals. Snakes have some physical properties to determine whether it is venomous or not. Heat pits on both sides of their noses, whether the head structure is triangular or has excessive and bright flake structure are some of these features. In this study, firstly, a data set related to the physical properties of snakes was generated. After this process, data mining classification algorithms J48 algorithm is used for the classification process for the detection of venomous snakes. Results show that high classification accuracy was obtained to predict snakes whether is venomous or not. Moreover, in case of the absence of some properties that give clear information about the snakes, it was investigated how is it possible to detect the venomous snakes. The results of this process were also compared with some algorithms often used in the literature. As a result of the study, it is concluded that what features are needed for the estimation of venomous snakes and that the best result can be estimated with J48 algorithm. © 2019 IEEE.