Analysing Crowd-Funding Campaign Success Factors using Machine Learning
Date Issued
February 11, 2021
Type
Μεταπτυχιακή Διπλωματική Εργασία
Abstract
In this past decade, an ever increasing number of businesses, small time creators, hobbyists, and entrepreneurs started using online crowd-funding platforms such as Kickstarter as a tool for funding their start-ups, kick start their ideas, or pre-sell their product to an audience in order to gauge interest. Consequently, there is a growing need for information that will allow these creators and entrepreneurs to plan, launch, and pivot their campaigns successfully. The aim of this work is to define the most important success factors for crowd funding campaigns using machine learning and deep learning techniques. Towards this, a web scraper for gathering Kickstarter campaign data was developed and as a result, a novel dataset on this research topic was created, including information on descriptive, financial and linguistic features, of around 2500 finished crowd-funding campaigns that had been run on Kickstarter. Machine learning experiments that included various algorithms such as Decision Trees and Support Vector Machines, but also Deep Learning Neural Networks were employed in an attempt to predict the successful funding of the campaigns of the dataset.
Subjects