Stock market is a specially and complicated nonlinear system,whose movement not only has its internal orderliness,but also affected by market,macroeconomic,and other noneconomic reasons. traditionally, methods just depend on linear analysis have many limits and difficult to forecast the exact results.the forecast precision in stock market is very important for one country's economy development and inverstors.Paper is based on the stock forecast , quantizes the part of stock's effect factors and syncretizes to the technologically analytical method in traditional securities, to adopt LM algorithm,and establish one stock forecast model based on the Neural Network (NN).Moreover,this paper will use the model to forecast Shanghai stock composite index,and estimate the model's validity.Firstly,this paper has analyzed the validity to forecast stock market and stock's factors of influence. it also analyses the stock forecast methods at home and abroad,and summarizes the new development in NN's forecast investigation.Secondly,this paper has expatiated basic theories about Data Mining , to discuss the BPNN model and structure,BP's study rules and algorithm which can improve BP calculate method.Finally, to filter the input variables by Data Mining algorithm, and use the updated BPNN algorithm which is LM algorithm to set up a model to forecast and emulational Shanghai stock composite index . the results show that BPNN can efficiently forecast the stock market, and has feasibility. |