Design Of Gsm Short Message Recorder And Its Forecast Function Implementation

As the development of wireless communication technology, the mobile phone has become an important and necessary tool. The GSM network is the most widely used network of mobile communication system. As the simplest and most convenient data communication mode, the short message service is highly popular mode by people. In the case of market commodity, such as newspaper and ticket order, table booking and so on, using the mobile short message to complete it will make people's life more convenience. The system of using SMS to transfer or report the sale data, will be a new merchandise planning.This dissertation designs the short message recorder based on the GSM for demand order, and makes research on the short-term demand forecast. This paper mainly contains the following research:First, in this paper analyze the sending and receiving mode of SMS, design the short message recorder controlled by MCS-51 P89V51 micro-controller, which based on the wireless module TC35i. The short message can be printed out by the micro-printer. In this paper designed a method of serial port expansion using the digitally controlled analog switch CD4052.Second, for enhancing the whole performance, the AYR micro-controller ATmega64 has been used to substitute the MCS-51, which can infinitely receive and send mobile message, the massage can be formatted and instantly print out. The device can realize facsimile transmission.Third, the short-time demand forecast of consumption has several problems, such as small quantity of data, difficult solution of nonlinear fit function and so on, which can be rightly solved by SVM method. In this paper the SVM prediction algorithm has been studied, and the example result has been compared with several traditional prediction algorithms. The parameter optimization of SVM model is grid search algorithms. The verification indicates that the SVM is more suitable for the short-time prediction.The GSM message recorder can automatically print out the receiving message, which is conveniently to look over or save the message. Using the function of SVM prediction, the user can make the sale foreground prediction according to the recorded data as selling, which will be usable for the corporation optimizing storage, and increasing benefit.