Analysis of succés factors for initial coin offerings and automatisation of whitepaper analysis using text-mining algorithms
Tutor / director / evaluatorBenichou, Laurent
Document typeMaster thesis
Rights accessOpen Access
Executives and visionaries have already explored many use cases of blockchain. However, the discovery of using cryptocurrencies, which are created by using blockchain, as means of crowdfunding in the form of initial coin offering (ICO) had a huge impact on shaping the way people invest in 2017. ICOs became one of the first use cases of blockchain that changed a highly regulated finance industry and they have been in high demand since the success of Ethereum project. Investing in ICOs allows small individual investors interested in tech to indirectly participate and get involved in revolutionary projects that aim to disrupt the state of the art of the current industries and set new social and economic standards. Linked to the mid and long-term success of the blockchain projects and its implementations into real practical applications, crypto-tokens can be used to purchase the services offered and in the case of high popularity and market trust, they can be also used for third-party purchases. The counterparts of investing in ICOs are strongly linked with its benefits. The myriad of opportunities, the inexistent barriers of entry and the early-stage of the involved projects make them extremely attractive for small investors. However, these investments offer little or no financial guarantees and are subject to important uncertainties about the feasibility of the business model and the background of the developing team. Many impostors blossomed to benefit from the enormous demand, especially over the course of 2017. In this paper, we put forward the points to be considered before investing in ICOs and supported our assumptions by conducting data analysis on 106 ICOs and surveys on 50 amateur investors. Based on our study, we realized that one of the major pain points is to analyse white papers, as it is both time consuming and amateur investors struggle to understand the maturity, the potential use cases and the value proposition of the projects. In this context, we developed a textmining algorithm with accuracy between 95%-100% by using R Studio and its natural language processing libraries to apply machine learning techniques on white papers in order to automatize the investment decision based on the scope of the project.