RESEARCH RADAR
“Predictors of NFT Prices: An Automated Machine Learning Approach” published in January 2023 in Journal of Global Information Management by Ilan Alon (MIB Visiting Faculty and University of Ariel, Israel), Vanessa P. G. Bretas (Dublin City University, Ireland), and Villi Katrih (Signex, Israel).
Introduction
The article analyzes market- and network-related aspects of non-fungible tokens (NFTs) pricing determinants using a comprehensive dataset from Signex.io and automated machine learning. Network factors, particularly Twitter and Discord members, were found to be the most important pricing determinants, driving NFT prices in a non-linear fashion. This study contributes to understanding NFT pricing by identifying the most impactful factors using AML.
Methodology
The authors used automated machine learning (AML) to identify the most relevant pricing determinants of NFTs, including market conditions, network factors, and NFT features. They used comprehensive data from Signex.io, a platform that tracks NFTs and their characteristics, from January 2022 to July 2022. The authors found that AML is superior to linear regressions as it examines many types of models simultaneously, explores non-linearities, and makes no assumptions about predictor distribution and stochastic properties.Results
The study considered the different dimensions of NFT pricing determinants, including market conditions, network factors, and NFT features. Feature associations matrix and impact showed the strength of the associations and identified the most important features driving model decisions, with Twitter and Discord members being the most important. The Partial Dependence charts illustrated the marginal effect of features on NFT pricing, revealing non-linearities in their relationship.Conclusions
This article discusses the factors that affect NFT prices in the context of Web3 and blockchain technologies. The authors found that social networks such as Twitter and Discord are relevant and important predictors of NFT prices, with bigger communities tending to command higher prices for their NFTs. However, the relationship between community size and NFT prices goes through peaks and troughs, indicating that tighter and more focused communities may provide a better price outcome. The use of bots in social media marketing to promote NFTs is also discussed, with Elon Musk contesting their effectiveness.The article also examines the relationship between Ether prices and NFT prices, finding that Ether prices and volume are not strong predictors of NFT prices as expected.
Authors
Ilan Alon Ph.D, Full Professor of Strategy and International Marketing;at Ariel University (Israel) and at University of Agder (Norway), Visiting Faculty MIB Trieste School of Management in Strategy and International Marketing. Advisor, Crypto Economist, and Board Member.Vanessa P. G. Bretas, Assistant Professor in Global Strategy Dublin City University, Ireland.
Villi Katrih, Signex, Israel, is a serial entrepreneur with over 15 years of experience in the tech sphere.
Useful links
Download full-text
