Date of Award
May 2023
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Management Science
First Advisor
Huimin Zhao
Committee Members
Sanjoy Ghose, Cheng Chen, Zhao Wang, Gang Chen
Keywords
Action roles, Comment semantics, Metafunctions, Multimodal analysis, Prediction, Reward-based crowdfunding
Abstract
While there has been a rapid growth of the reward-based crowdfunding market, successfully achieving the target amount of a project is a big challenge. Entrepreneurs are eager to know the prospects of their campaigns. Investors also want to fund promising campaigns with high quality and low uncertainty. Therefore, effectively predicting the success of a project throughout the fundraising period is a crucial task for both entrepreneurs and investors. On the one hand, it provides guidance to entrepreneurs about project progress and potential, helping them adjust their campaigns in time. On the other hand, it helps investors manage their funding risks and reduce opportunity costs. However, most of the existing research has been aimed to explore the determinants of fundraising success, but much less attention has been paid to the success prediction problem. We explore the crowdfunding success prediction problem from two perspectives in the following two essays.In Essay 1, we mine semantic features from comments to improve fundraising success prediction. More and more participants share and discuss facts and opinions about projects by posting comments, which can influence investors’ funding decisions. Previous studies have mainly focused on quantity, sentiment, and linguistic features of comments, largely overlooking the value of semantic features, in predicting fundraising success. Rooted in information asymmetry and herding behavior theories, we posit that discovering semantic signals from comments and distinguishing actor roles will benefit fundraising success prediction. We propose a framework with novel latent semantic features of comments. Empirical evaluation using data from a prominent platform demonstrates the utility of the framework and reveals interesting patterns in the dynamic predictive effects of semantic features for different actor roles. In Essay 2, we apply features from multimodal data (texts, images, and videos) to improve fundraising success prediction. With the development in artificial intelligence and big data, multimodality has become one of the popular research areas of IS since multiple modalities can provide complementary information and improve the performance of the overall decision-making process. However, there is a lack of research providing a comprehensive investigation of multimodal data in crowdfunding. To gain a comprehensive review of linguistic and visual features of multimodality in crowdfunding, we propose a framework built on theories of Halliday’s metafunctions framework of languages (1985), Kress and Van Leeuwen’s functional visual design (1996), and Royce’s intersemiotic complementarity of languages and visual images (1998) to explore relevant features representing the ideational, interpersonal, and textual metafunctions of multimodal data in crowdfunding. We have conducted several experiments to study the effectiveness of each metafunction, each modality, and their interactions in predicting crowdfunding success.
Recommended Citation
Bao, Liqian, "Predicting Fundraising Success in Reward-Based Crowdfunding" (2023). Theses and Dissertations. 3122.
https://dc.uwm.edu/etd/3122