The Research On The Model And Algorithm Of Forest Disease And Pest Forecast Based On The Rough Set Theory

As the most precious strategic resources of human society development, is constantly suffer from huge loss of various disasters. However in these major forest resources, the frequent biological disaster such as forest diseases and pests has been to limit the forestry development. With the artificial afforestation area continues to grow, forest diseases and pests is always in upward trend. Therefore, we should take active measures to reduce the loss and avoid the damage caused by forest pest and disease.With the wisdom of forestry increasing demands continuously, in order to make the modern forestry informationization, networking and intelligent, the rough set theory, expert system, network technology, modern information technology and experts in the field of forest diseases and pests diagnosis, prediction and control experience and technology combine, application in forest diseases and pests forecast, diagnosis and prevention, so as to reduce harm caurse by forest diseases and pests. It promotes the modernization, scientific of the management of forest diseases and pests. And it makes the realization of the development with the type of span informatization. This plays an important role to advancing demonstration areas of forest diseases and pests monitoring, early warning, prevention, control system and mechanism innovation.In order to forecasting of forest diseases and pests occurrence accurately, this paper analyzed the key factors of effect the forest disease and pests occurrence, such as forest structure, climate change, biological factors, soil factors and human activities, and make a detailed analysis of the factors of the temperature, humidity, rainfall, light, wind, and the influence of thess effects on forest disease and pestd occurrence and development. This paper cited the temperature and solar radiation simulation model and the humidity and precipitation simulation model, the simulation model of the seasonal weather can be obtained within the daily maximum temperature, minimum temperature, precipitation, meteorological factor of the simulated data, which is conducive to make the forecast accuracyIn order to timely control of forest diseases and pest population vari(?)tion of forest diseases and pests, the future situation and growth tendency to make scientific, accurate forecast, so as to give forest diseases and pests control measures, reduction of forest pests and diseases damage and loss. This paper applises developmental schedule forecasting method, effective accumulated temperature prediction method, regression analysis on demonstrative area common forest pest occurrence period forecast, and the establishment of the corresponding prediction model, application of effective base prediction method, mathematical and statistical method of demonstration area common forest pest occurrence quantity forecast, and the establishment of the corresponding prediction model, using Markov process analysis, demonstration area of common forest disease severity prediction using grey model for demonstration area, common forest disease process simulation and onset time prediction by using the non linear model, demonstration area of common forest disease stage forecasting, gray model analysis on utilization, demonstration common area of forest disease index predictioy.Forest diseases and pests forecasting expert system includes large quantity knowledge. In order to mining potentially valuable information from the mass data quickly, accurately and uses for the prediction of forest pests and diseases. This paper uses the rough set theory and applises it to the process of the prediction of forest pests and diseases. Through the forecast data are collected, complete and discretization, put forward a kind of improved attribute reduction algorithm based on discernibility matrix. Based on this algorithm on forest diseases and pests forecasting condition attribute set is reduced, thereby to generate rules extraction, obtains forest diseases and pests forecasting model based on a new rough set theory.This paper adopts three layers of B/S structure and based on the J2EE standard development mode for the establishment of forest diseases and pests forecasting expert system. This model can make the separation of the performance of the front and the application logic, data storage phase. It makes the whole system structure clear and flexible, convenient deployment and expansion through the component development and deployment strategy. Business processing with Jsp+Javabean mode and accessing data resources from data layer through the way of JDBC. Forest diseases and pests forecasting expert system based on demonstration area common forest diseases and pests occurrence period and amount of prediction and assessment of loss. In addition to this, the expert system also includes species information, diagnosis of diseases and pests, plant diseases and pests, diseases and pests query and other auxiliary module.In summary, this paper's main contribution is as follows:(1) Rough set theory and artificial intelligence technology was first used in the field of forestry diseases and pests;(2) Proposed the model of forestry diseases and pests forecasting in the demonstration. Through the model validation and accuracy analysis, the accuracy of the result is high, and forecast results with the actual situation are match;(3) Proposed an improvement attribute reduction algorithm based on discernibility matrix. Reduced conditional attribute set of forestry diseases and pests forecasting based on the algorithm. Through the extraction and reduction of generation rules, obtained a new f forestry diseases and pests forecasting model based on rough set theory. By verifying, the model has achieved a good result. The research results of this paper can provides the theoretical guidance and technical support for the vast number of users and farmers to prevent the forest diseases and pests. And provides a complete sample for the digital forestry construction. It has the certain instruction significance and reference value on the expert system of forecast and prediction of agricultural pests and diseases.