Similarity analysis in the context of ICO whitepapers

Details

Supervisors

James Thewissen

Faculty

Louvain School of Management

Degree label

Master [120] en sciences de gestion, à finalité spécialisée

Abstract

The basis of the paper is to study the innovation aspect of initial coin offerings (ICOs), more specifically whitepapers, and to see how innovation has an impact on fundraising and success. We filtered these text documents into different topics using Topic Modeling to analyse the similarity of these documents. The filtering was done with Latent Dirichlet Allocation (LDA) and the categorisation of topics was done by human judgement. Using this method, we were able to analyse the similarity of the whitepapers and the results show that the similarity of the whitepapers has a positive impact on fundraising provided that the similarity is not too great. In other words, there is a curvilinear relationship between similarity and fundraising. On the other hand, a very low textual similarity implies a higher chance of encountering a potentially fraudulent ICO, thus fraudulent projects bring innovation in an industry with few competitors to attract investors looking for rapid profit.