Font Size: a A A

Design And Implementation Of Fund Selection System Based On Machine Learning

Posted on:2019-03-14Degree:MasterType:Thesis
Country:ChinaCandidate:M Y LiuFull Text:PDF
GTID:2428330545952279Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of the securities industry,stock trading has become more and more popular among investors,securities companies,fund companies and SFC.The effective management of listed companies can not only arouse the enthusiasm of investors,but also guarantee the industrial development of enterprises,thus realizing the maximization of enterprise interests.The fund optimization system of this project mainly relies on the relevant fund data,and finally realizes the investment analysis and decision-making.It integrates the cur-rent advanced machine learning engineering design concept and the more mature finan-cial system knowledge.According to the actual needs,the project is divided into four modules:data processing module,feature selection module,machine learning module and fund optimization module.Data processing module mainly deals with data collected by third-party data providers;The feature selection module mainly focuses on the char-acteristics of the data processing data and the statistical formula.The machine learning module is mainly the realization of the machine learning algorithm after the selection.The optimal module of the fund uses the predictive value of the machine learning mod-ule to give the strategy according to the customer's request.The data processing module,feature selection module and machine learning module of these four modules are com-pletely analyzed,designed,implemented and tested by me.The fund optimization mod-ule is only responsible for background analysis,design and implementation,and the front-end design,implementation and testing are all done by others in the team.The main technology of the system is machine learning,which is combined with fund optimization to simulate investment and other relevant functions,and provides new ideas for fund manager decision-making.In the system testing phase,this topic will design test cases for feature engineering,machine learning and fund optimization.I hope that the stability of the control algo-rithm can be obtained through the system test.This system is an experiment in the current quantitative environment.The current quantization is for futures,and there are few targets for stocks or even funds.Although the result of this system can not be separated from the people's control,but it actually plays a positive role in the decision-making process.
Keywords/Search Tags:Machine learning, finance, data processing
PDF Full Text Request
Related items