| Quantitative trading analysis has attracted more and more attention.It has become a hot field of computer application.Nowadays,people do not like investing with experience and feelings,and more hope that through scientific data analysis for reasonable allocating funds.Stock is a popular investment product.And a lot of people want to try stock data analysis and get the stock price change rule.As the enlargement of the quantitative analysis demand,open quantitative investment platform come out to provide the corresponding function.But there still exists certain disadvantages of the open quantitative analysis platform.Especially,the platform only provide the daily trading data,which will be not enough for quantitative strategy design and will make the accuracy of back not so high.There have some optimization space between the trend judgement and the actual transaction.In this thesis,we starts from introducing the function and usage of open quantitative investment platform and present that we must turn to finer-grained data.But fine-grained stock data,especially tick-by-tick trading data,is a threshold for the beginners,which makes them do not know how to start while looking at the treasures.They can not process the data efficiently.Further more,this thesis will discuss the disadvantage of old back test model and design a new one to support fine-grained stock data operation.With the combination use of coarse-grained trading data and finer-grained trading data,we can produce multidimensional index,which will extend the tend judgment to the master of interaction point.In the new back test model,we can process the data in advance and optimize the storage/accessing part.Columnar has been chosen for the trading datastorage.On this basis,difference approximation algorithm is design to reduce the amount of trading data.As a result,the amount of calculation will be reduced.In-memory compression technique has also been adopted to improve the speed of data processing.The data file will be also saved as the format in memory,which will save much disk space and improve the speed of file reading.To sum up,stock trading data quantitative analysis is discussed in this thesis and a new back test model is presented to solve the existing disadvantages.The efficiency of data storage and accessing is improved by designing the storage algorithm and optimizing the file reading methods.As a result,we can use the multidimensional index to enrich the quantitative strategy designing.,which can improve the accuracy and benefit while decreasing the risk.It shows the application of computer technology and ideology in stock trading data quantitative analysis,storing and accessing. |
