Research On Data Mining Technology For Semiconductor Device

Solar cells is the use of the photovoltaic core device, but also a representative of thesemiconductor device which is based on the equivalent circuit analysis and its batteryperformance parameters are analyzed to predict the direction of the solar cell is animportant research.Firstly,this paper introduces the basic knowledge of data mining techniques andanalyzes the equivalent circuit of the traditional silicon solar cells and thus to serve as thebasis for further electrical parameters of the solar curve fitting, using third-party software1stOpt with Matlab and Walmer-Butler kinetic parameter estimation methodssimultaneously, highlighting the reliability and accuracy by comparing Matlab method.Secondly,this paper introduces the BP neural network data mining algorithms tocalculate the basic structure and processes, and the use of Matlab software combines BPneural network algorithm parameters to predict solar written procedures. Then use basedon BP neural network algorithm to predict a set of procedures to predict the performanceparameters of outdoor solar NSL U.S. data obtained from the National Laboratory. Afterthe results of verification, use the higher precision of the IV curve written thesis based onBP neural network algorithm to estimate the parameters of the fitting procedure.Finally,realized by Visual C#and Matlab parameter prediction function andimplement the application of visualization, C#advantage is that you can develop softwarecan be developed. Applications exe format, no longer dependent on the source code, youcan directly file packet copy to be used from any computer, restrictive become very small.