基于GA-BP神经网络模型的奖励式众筹融资结果预测研究

发布时间：2018-05-26 13:10

本文选题：众筹 + 遗传算法　；参考：《上海师范大学》2017年硕士论文

【摘要】：近年来,互联网金融正以一种蓬勃发展的势头影响着各行各业,由于其具有方便、快捷的特点,对人们的生活有着深远的影响。当前,社会已经形成了“互联网+”的大背景,互联网金融不仅帮助完善线下业务,在线上业务的模式上也有很多创新。作为互联网金融的产物,众筹模式利用互联网帮助某一特定项目进行融资,在形式与实践上都与以往的投融资模式有着很大的区别。由于众筹网站的便捷性和广泛性,为当前中小企业融资难的问题提供了一个很好的思路,同时,也为民众提供了直接参与金融市场的渠道,丰富了民众的投资选择,有利于实现民间资本与中小企业的高效对接,缓解资本市场资金紧缺的压力。因此研究众筹模式对于当前我国经济有着重要的意义。针对当前我国众筹模式以奖励式众筹为主的特点,本文以奖励式众筹作为研究对象,正文研究主要分为五个部分。第一部分,本文对当前我国奖励式众筹的融资背景进行了分析,并指明了本文的研究意义以及研究思路;第二部分,本文对前辈学者的研究加以总结并详细介绍了本文研究中所涉及的理论;第三部分,对于本文研究对象奖励式众筹的要素以及众筹模式做了重点介绍并从信息不对称性理论、期望理论和羊群效应理论出发分析出当前影响众筹融资成功的五个突出影响因素,分别是项目经济性因素、能力信任因素、信息质量因素、项目规模因素以及社会网络因素;第四部分,本文根据五大影响因素结合“摩点网”的相关众筹项目数据选取出相应的九大变量指标,分别为项目人微博粉丝数、项目人历史支持过的项目数、项目人历史发起的项目数、项目是否具有视频介绍、项目更新次数、筹资时间、项目最小单位的筹资额、回报方式数目以及众筹目标筹资额。利用归一化对数据进行标准化处理,代入GA-BP神经网络构建出一个奖励式众筹融资预测评价模型,并通过实验测试验证模型的精确性;最后,本文对于全文成果加以总结并给出相关建议。本文的创新之处在于将理论研究与实证分析相结合,利用遗传算法和BP神经网络原理构建GA-BP预测模型,对于当前我国的奖励式众筹模式进行深度地解读,因此可以为众筹项目方获取众筹成功提供一定的指导建议。
[Abstract]:In recent years, Internet finance is affecting all kinds of industries with a vigorous development momentum, because of its convenient and fast characteristics, it has a profound impact on people's lives. At present, the society has formed the "Internet" background, Internet finance not only helps to improve offline business, online business model also has a lot of innovation. As a product of Internet finance, crowdfunding mode uses the Internet to help a particular project to finance, which is different from the previous investment and financing mode in form and practice. Because of the convenience and extensiveness of crowdfunding websites, it provides a good way of thinking for the current problem of financing difficulties for small and medium-sized enterprises. At the same time, it also provides a channel for the public to directly participate in the financial market and enriches the investment choices of the people. It is beneficial to realize the efficient docking between private capital and small and medium-sized enterprises and relieve the pressure of capital market capital shortage. Therefore, the study of crowdfunding mode for the current economy has an important significance. In view of the characteristics of the current crowdfunding model in our country, this paper takes the reward crowdfunding as the research object, the main body of the study is divided into five parts. In the first part, the paper analyzes the financing background of incentive crowdfunding in our country, and points out the significance of the research and the research ideas; the second part, This paper summarizes the previous scholars' research and introduces the theories involved in this research in detail. The third part focuses on the research object, the elements of reward crowdfunding and the model of crowdfunding, and introduces the theory of information asymmetry from the point of view of the theory of information asymmetry. Based on expectation theory and herding theory, this paper analyzes the five outstanding factors that affect the success of crowdfunding, namely, project economy, ability and trust, information quality, project scale and social network. In the fourth part, according to the five influential factors and the relevant crowdfunding project data of "friction Point Network", this paper selects the corresponding nine variables, which are the number of Weibo fans and the number of projects supported by the project owner in history. The number of projects initiated by the project person, whether the project has a video introduction, the number of project updates, the time of raising funds, the amount of funding of the smallest unit of the project, the number of returns and the amount of funds raised by the crowdfunding target. Standardized processing of data is used to construct a model for predicting and evaluating reward crowdfunding by means of GA-BP neural network, and the accuracy of the model is verified by experimental tests. This article summarizes the full text achievement and gives the related suggestion. The innovation of this paper lies in the combination of theoretical research and empirical analysis, using genetic algorithm and BP neural network principle to construct GA-BP prediction model, and to deeply interpret the reward crowdfunding model in our country. Therefore, we can provide certain guidance suggestions for public-funded projects to obtain crowdfunding success.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51;F724.6

【参考文献】