Paper Title
Application of Artificial Neural Network for Predicting the Success of Non-Fungible Token Creation

Abstract
This research investigates the application of Artificial Neural Network (ANN) for predicting the success of Non-Fungible Token (NFT) creation. The data are collected from the OpenSea platform, which is widely accepted as a marketplace for NFT trading. The trading volume represents the success of NFT creation. The predictors include the number of NFTs, number of owners, floor price, percentage of the listed NFTs, and percentage of unique owners. Compared to the Probit model, the ANN shows superior predictive ability. The number of NFTs, the statistically insignificant predictor from the Probit model, could be used to improve the ANN's predictive ability. The result reports that the use of non-linear programming and learning ability bring about better performance of the ANN over the traditional statistic model. Furthermore, the ANN has a superior process for finding the relationship between the independent and dependent variables. The results also prove that ANN could be a useful method for selecting the predictors. Hence, the ANN is a helpful prediction technique, especially for the new asset with unclear predictors. Keywords - Artificial Neural Network, Probit, Marketplace, Non-Linear Programming, Non-Fungible Token, Platform.