Research on Vespa Mandarinia's Invasion in the State of Washington

Abstract

In recent years, a new invasive species, Vespa Mandarinia has become a problem for the State of Washington, the U.S.A. and regions near it. In this research, we used Kernel Density Estimation, Natural Language Processing and Convolution Neural Network(CNN) to evaluate the geographical and textual data of civilian reports and in what way we can detect new invasion cases without a professional's presence. The results of the research show that we could perform prediction with data available but it could be biased. However, image classification techniques based on CNN could be an incentive for more data input, therefore lead to a better simulation and estimation. The method of our research also indicates that these techniques may have great applicability to other invasive species.