Fair NFTs evaluation based on historical sales, market data and NFTs metadata

Abstract: Blockchain technology is rapidly growing and with it, the opportunities it brings. Non-fungible tokens (NFTs) are a type of token that represents unique data that can be owned and traded on a blockchain. The NFT market is still in its early stages, making it difficult for newcomers to enter without a solid background and knowledge of the technology. In this study, we aimed to discover if the new information provided by blockchain and NFTs can be used to build better estimation models. We focused on the availability of every single transaction and the attributes attached to NFTs. Our results showed that the availability of each transaction improves the accuracy slightly, while the attributes attached to the NFT greatly improve accuracy. We are confident that with more data over time, the accuracy will continue to improve mainly if the transactions are able to cover almost all the NFTs inside a collection because all the different combinations of attributes will be analyzed.