Author(s)

Feier Yang

Corresponding Author

Feier Yang

Abstract

This article presents a dandelion spatial propagation model and a time series prediction model to elucidate the seed propagation mechanism and predict future propagation ranges of dandelion. Additionally, a species feature selection model and an invasion degree evaluation model based on XGBoost have been developed to identify key factors influencing species invasion and to assess the areas most affected by such invasions. The spatial propagation model incorporates climate data, including air humidity, temperature, dandelion density, wind speed, and direction, to simulate the effects of wind speed and air humidity on seed dispersal in various directions. The findings indicate that the smallest transmission range occurs in January and December, averaging 1.16 meters, attributed to low wind speeds and high air humidity levels. The study also pinpoints four significant factors impacting invasive species: species diversity, destruction rate, transmission rate, population density, and the environmental pollution rate. The models have been applied to determine that Amaranthus spinosus exhibits the highest invasion degree in Jiangxi, while water hyacinth is the most invasive species in Guangxi.

Keywords

Space communication; Time series; XGBoost; Feature selection; Intrusion level