Machine Learning Crowdfunding

Machine Learning Crowdfunding

Evangelos Katsamakas (Gabelli School of Business, Fordham University, New York, USA) and Hao Sun (Gabelli School of Business, Fordham University, New York USA)
Copyright: © 2020 |Pages: 11
DOI: 10.4018/IJKBO.2020040101

Abstract

Crowdfunding is a novel and important economic mechanism for funding projects and promoting innovation in the digital economy. This article explores most recent structured and unstructured data from a crowdfunding platform. It provides an in-depth exploration of the data using text analytics techniques, such as sentiment analysis and topic modeling. It uses novel natural language processing to represent project descriptions, and evaluates machine learning models, including neural network models, to predict project fundraising success. It discusses the findings of the performance evaluation, and summarizes lessons for crowdfunding platforms and their users.
Article Preview
Top

Introduction

Innovation is the driving engine of economic growth and prosperity. However, innovators need funding to realize their innovations and bring them to market. In the past, a variety of actors, such as banks and venture capitalists, provided this funding. Crowdfunding, is a novel decentralized approach to funding projects, facilitated by digital crowdfunding platforms.

Crowdfunding is the practice of funding a project or venture by raising money, typically small contributions, from a large number of people, using an Internet-based crowdfunding platform (Belleflamme, Lambert, & Schwienbacher, 2013, 2014). Crowdfunding has been used to fund a wide range of for-profit entrepreneurial ventures such as artistic and creative projects, social entrepreneurship projects, medical expenses, travel and many others.

Crowdfunding is a type of fintech (financial technology), a set of technologies that play a key role in transforming or disrupting various finance functions and processes. Moreover, crowdfunding is a form of alternative finance. Social lending, or P2P lending, is another example of alternative finance. Crowdfunding is also a form of crowdsourcing, which in turn evolved from open sourcing in software industry (Economides & Katsamakas, 2006; Katsamakas & Georgantzas, 2010). The literature identifies four types of crowdfunding: rewards-based, lending, donations and equity. This article focuses on rewards-based crowdfunding.

The modern rewards-based crowdfunding model is generally based on three types of actors: the project initiator who proposes the project to be funded, individuals who support the project, and a mediating organization (the “platform”) that brings the parties together.

A crowdfunding platform is an instance of the phenomenal growth of platform business model in digital economy (Bakos & Katsamakas, 2008; Eisenmann, Parker, & Van Alstyne, 2006; Parker & Van Alstyne, 2005). Project creators may receive several potential benefits from using a crowdfunding platform:

  • The innovator is able to test the market demand for his proposed project;

  • The innovator receives early feedback from the crowd;

  • The innovator develops a relationship with prospective customers and this relationship can be beneficial over time;

  • Innovation risk is reduced.

In this research, we analyze data collected from Indiegogo, a popular international crowdfunding platform founded in 2008 (see https://www.indiegogo.com/). About ten million people visit the site each month. Indiegogo allows people to solicit funds for an idea, charity, or start-up business, and it charges a small percentage fee on each contribution. The site runs on a rewards-based system, meaning that users who are willing to fund a project or product will receive a reward or gift, rather than an equity stake in a company. Following changes in Security and Exchange Commission rules in 2016, Indiegogo offered an equity crowdfunding service, as well, but this it outside the scope of this article.

With the increasing number of projects competing for funding in a crowdfunding platform, it is important for project initiators to present their project in a most attractive way to raise the funds they need. In this research, textual description is utilized to understand whether text can make a project more successful in attracting investment.

This article uses text mining and natural language processing (NLP) to explore and model text data, and machine learning to predict crowdfunding project success. A main contribution is that we evaluate a number of different techniques in context of crowdfunding, especially recently developed techniques for text representation and neural network (deep learning) techniques for classification.

The rest of the article is structured as follows. Section 2 discusses research on crowdfunding. Methods of data collection and exploration are presented in Section 3. Section 4 presents machine learning experiments and results. Discussion and conclusions are in section 5.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 10: 4 Issues (2020): 3 Released, 1 Forthcoming
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing