Design And Implementation Of Stock Information System With Respect To The Big Data Environment

Our country through a relatively short time in the stock market has reached the developed countries in the struggle of nearly the size of the hundreds of years.But such rapid growth also exist many problems to be solved,including the problem of the large amount of data in the stock market.Firstly this paper introduces the research background and research significance of the stock information system,through the common stock prediction methods are introduced both at home and abroad,real-time understand shortcomings existing in current stock prediction process and researchers have put forward a series of advantages;Secondly,this article uses data mining to the three kinds of methods are introduced in detail,mainly introduced the decision tree classification algorithm,association rule algorithm and neural network algorithm,the Map Reduce software architecture and apply method is also introduced;Next,the stock information system requirements analysis and system design,including the analysis of user management,transaction management,stock management,query functions and text messages warning management five modules such as use case analysis,and these modules in detail design,and to establish the system of the whole architecture;Moreover,three methods of data mining in the application of stock in detail,including the application of decision tree classification algorithm in the stock,the application of association rules in the stock and the neural network model in the application of stock and so on three parts,then the Map Reduce architecture is used to improve the efficiency of algorithm,and through the experiment results show that the method has certain rationality and practicability;Finally,in this paper,the stock information system of each interface implemented and tested.By completing the whole stock information system,in this paper,the main contributions are as follows:(1)develop a satisfy the users of different professional stock information system,to join the data mining algorithm makes the stock information system more practical;(2)the decision tree classification to the stock,and analyze the classification result and model evaluation;Then use association rules analysis,found that the stock plates,the principle of the strong correlation between regional.Cycle of data by choosing stocks,the decision tree classification method was used to construct aspecific field of stock classification model,according to the result of the experiment of classification and analysis,found that the method has high accuracy in the stock prediction.(3)using association rule in the classic Apriori algorithm is applied to the security analysis,so as to dig up has guiding significance to the rules of the stock prediction.(4)through the application of artificial neural network model to apply,the stock index is mainly based on the market the highest point of the predicted curve compared with real peak index curve concluded that BP network to predict the stock inventory has the advantages of good effect,high precision of data fitting,and using the Map Reduce to improve the efficiency of BP algorithm,experiments shows all the methods are suitable to the stock market data dealing.