The Researching Of Some Forecast Methods In Precision Agricultural

Precision agriculture, which can precisely adjust managements about soil andplants according to the practical circumstance by making use of geographicalinformation systems , global positioning system and remote sensing technology, hasbeen optimizing the application of agricultural investments and maximizes theagriculture output and economy benefit, it protects agricultural resources such asenvironment and soil as well as. Precision agriculture, being comprised of interactedsystems of precise seeds engineering, precise planting, precise fertilizing, preciseirrigating, flexible plants regulating and precise harvesting, is a complex system whichis driven by data and in the process the digging data is most important. The mainlytechnique of digging data technology is how to select and setup the mathematicmodel. In the precision agriculture, experts and artificial intelligence data analyzingsystems sustain the decision‐making system. Being the precondition and groundworkof making decision, the precision of speculating and forecasting data dominates thefeasibility of the decision. So the speculating and forecasting data is the focus in themaking decision and sustaining systems in the precision agriculture, but it is still ablank in our counties precision agriculture.None of the alone implement can use every available information in practicalforecasting data. It is always happened that different result will be given whendifferent method is taken. The better precision can be obtained by using appropriatecombination of several different speculating implements. Therefore, only whenproper mathematic model is selected flexibly and effectively according to differentproperties of different studying object, can the precise and reliable predicts be makesure. There are many factors which have an influence on the forecasting in thepractical analyzing process. The factors are interacted and associated with each other,so it is normal that the problems of forecasting always are concerned plural entities.With traditional math people can resolve problems of linear multifactor functionsand some simple problems of nonlinear multifactor functions, but for most of thenonlinear multifactor problems, the complicated or inaccurate forecasting resultswould be obtained when using traditional math tools, which make forecastunavailable. With the appearance of mathematic model base on artificial neuralnetworks, for example, BP neural networks, radial factors functions neural networks,broad return neural networks, it is possible to improve and optimize the forecasting and speculating data in the precision agriculture. For artificial neural networkspossess excellent ability of tolerating mistakes, it can carry out complicatednonlinear mapping easily. And because of its great ability of computing, it has beenbeing applied widely in computer, automation and artificial intelligence and by whichmany difficulties that can not be resolved by traditional implements have beenresolved. It is certainly that artificial neural network mathematic implement have itslimits and shortcomings for it is still a new coming thing, but with the growing of thetechnology, it will make great change for us.MATLAB, a large‐scale computer assistant software which possesses a largeamount of calculating and drawing tools in its toolbox and has a very compactpeople‐computer interface, ensures the complicated calculating being carried out.Integrating the ANN and MATLAB will make analyzing and calculating systemscontaining large amount of variables become possible, furthermore, the complicateddifficulty of forecasting will be resolved.