Track: Undergraduate Student Paper Competition

Abstract

NFT or non-fungible tokens are online certificates of ownership that can be traded, based on data units stored in digital ledgers (ledgers) belonging to blockchain technology. Non-fungible means that it cannot be exchanged and is unique. NFT has been around since 2014. But now, it is increasingly being considered as a fairly practical method for trading digital artwork or art. To buy NFT assets, it requires special coins in the form of NFT coins, which consist of various types, such as mana coins, sand, axs and other NFT coins. The NFT coins are used to process NFT purchase transactions. The movement of NFT coins over time is relatively erratic and uncertain. This NFT coin price prediction will be very useful for investors to know how the investment flow of each price works, because the price of each NFT coin will change from time to time. Through the literature study stage, interviews and viewing daily NFT coin price data where the attributes used are date, open, high, low, close volume, and marketcap. Based on the experimental KNN (K-Nearest Neighbors) method on a data set using the parameter values of K 3, 5, and 7, it can be said that the KNN model that has the best accuracy is KNN with a value of K=7 with an accuracy of 65% . The greater the value, the greater the Mean Square Error and Mean Absolute Error values. The research conducted predicts the Close value of the NFT coin price in a period of 1 day.

Keyword: Blockchain, NFT (non-fungible token), KNN (K-Nearest Neighbours)