Forecasting NFT Prices on Web3 Blockchain Using Machine Learning to Provide SAAS NFT Collectors

Abstract

Non-Fungible Tokens (NFTs) are one-of-a-kind digital items with static or continuous visual and audio content. NFTs digitally represent any assets that may hold photos, gifs, audio, videos, or any other data-based storable material. These assets may come under a variety of asset groups, including art, in-game goods, and entertainment collecting units. What makes them appealing is their exclusivity, in the sense that each NFT is unique to itself, and ownership is determined by a digital certificate. In the first half of 2021, NFT sales totaled more than a billion. The NFT Software as a service (SAAS) based system is a one-of-a-kind offering and concept for thinking outside the box and presenting intellectuals and creative treasures and exhibiting these objects to ensure the security and integrity of digital assets. The existence of core decentralized networks allows for unrestricted access to this material as well as further analysis. Based on the Web3 Blockchain technology, these assets may be traded and represent next-generation ownership. In this paper, Adaptive Improved Convolutional Neural Networks (AICNN) are used to forecast NFT to provide a SAAS NFT collector. We also introduce Tree-seed Chaotic Atom Search Optimization (TSC-ASO) algorithm to optimize the forecasting process. The proposed method of NFT price forecasting is evaluated and compared with the existing forecasting methods. To produce an accurate report for NFT price forecasting, the proposed method will be effective.