Analysis of Hornet Forecast Model based on Fuzzy Theory

Abstract

The 14 Positive ID the paperre arranged in time order and the GM Model was used to predict the range of propagation, and the results of the prediction the paper re obtained as follows: from 48.92 to 49.05 in length and from -122.47 to -122.55 in latitude in 2021. There is a distance of 30 km betthe paperen the predicted results and the initial point where the presence of hornets was confirmed. The average relative error is less than 0.01, so the model prediction accuracy is good. Since the life cycle of hornets is very related to seasons, the time is converted into seasons and then One-Hot-Encoding of seasons; the TFIDF Algorithm is used to calculate the importance of each Note to replace the original Notes. The SMOTE Method used in this paper to fill the Positive ID minority class sample leads to the proliferation of Vespa mandarinia seriously endangering the local ecology, so the SMOTE Method used in this paper to fill the Positive ID minority class sample. The models all seek to maximize the recall of a few classes of Positive ID. After model testing our models are all excellent in identifying pests accurately, as evidenced by the ROC (with Positive ID as a positive example) curve and AUC =0.99.