| Wheat scab is one of important diseases in the world.Scab affects the yield and quality of wheat and endangers human life and health,so it is very important to predict and prevent the occurrence of scab in advance.The occurrence of wheat scab is affected by many factors such as temperature,humidity and wheat variety,so it is difficult to predict the accuracy.In this paper,with wheat in Shandong province as the research object,the meteorological data and field disease rate of wheat scab in Dezhou,Liaocheng,Heze and Jining of Shandong province during the years of scab occurrence were summarized and screened,and the prediction model was established based on Python to study the probability of scab occurrence.The specific work contents and results included:(1)Data preprocessing.In this paper,the temperature and humidity data,which have important influence on the occurrence of scab in wheat,and the field rate of scab in some urban areas of Shandong Province were collected.The correlation analysis between the available data after grouping and the diseased field rate was carried out to obtain the meteorological data variable which could be used as the independent variable factor.(2)Model One--linear regression model.A prediction model of wheat scab in Shandong province based on multiple linear regression was established.Several meteorological variables and cross-term factors with strong correlation with field rate of wheat scab were analyzed by regression,and the independent variable with the strongest correlation was selected to establish the prediction model.(3)Model two--Random forest or decision tree.A prediction model of wheat scab in Shandong province based on random forest regression was established.The model was constructed by using the data of Dezhou,and an optimal model was obtained by adjusting the optimization parameters,which could accurately predict the field rate of scab in Dezhou.Jining and other cities to verify the results of the accuracy of more than 80%.(4)Model 3--Support vector machine regression.A prediction model of wheat scab in Shandong province based on SVR(support vector machine regression)was established.The algorithm and principle of support vector machine and support vector machine regression are introduced.A loss function is introduced to input the data of Texas as the training set,and other data as the test set for verification.It was found that there was an important correlation between wheat scab and humidity and temperature in April and May,which was consistent with the actual situation of scab outbreak in May.It can be concluded that random forest has better prediction of wheat scab and can be used as a reliable model for wheat scab epidemic prediction in Shandong province. |
