Metaverse token price forecasting using artificial neural networks (ANNs) and Adaptive neural fuzzy inference system (ANFIS)

Abstract

This study is about the metaverse environment which has recently heard a lot in life. Although many individuals and institutions are interested in metaverse, it is an imaginary future space, and the conceptual framework is not fully drawn. The metaverse is a mix of augmented, virtual, and mixed reality technologies that are predicted to affect our lives over the next decade including our money, possessions, and ownership. So, we examined to forecast metaverse token price using ANN and ANFIS methods. The market value of the five metaverse token firms: opening price, highest value, lowest value, closing price, volume value, and variables, is evaluated between October 17, 2017, and December 15, 2022, by YSA and ANFIS. The ANN method training was carried out using ten hidden layers and the Levenberg–Marquardt algorithm. In the ANFIS method, training was carried out with 100 iterations in a triple mesh structure. At the end of the method training, the training performed with the ANN method was 99.3% successful, and the ANFIS method was 99.1% successful. It is concluded that the model fitness of the R2 values of ANN and ANFIS methods is appropriate at 99.3 and 98.7%, respectively. As a result, ANN and ANFIS methods can be used for the prediction of metaverse token prices for the estimation of financial instruments. ANN and ANFIS are advanced tools for predicting metaverse token prices, with ANFIS having unique features like fuzzy logic. However, using only basic price data is not enough for precise predictions. While the EMA and SMA have less impact, gold (XAU), BTC, ETH, US dollar (USD), Chinese Yuan (CNY), and Brent oil (BRT) are observed to have a moderate impact on determining the market values of metaverse prices. The use of XAU and ETH prices in both ANN and ANFIS methods gives successful results. Especially, we have achieved favorable outcomes when employing the ANN method for analyzing EMA and BTC values. Additionally, we have obtained valuable and successful results by utilizing the ANFIS method for analyzing BRT. Using only opening, highest, lowest, closing prices, and volume values and USD, CNY, BRT, and SMA prices has not demonstrated usefulness in attaining favorable results. The findings of this research indicate that the utilization of the metaverse world has the potential to enhance learning capabilities and motivation, as well as make significant contributions to industrial production and the field of health. However, it is important to note that the existing body of research on the metaverse world remains limited in scope and depth. Users often hold preconceived notions and biases regarding ethical concerns associated with the metaverse world. Additionally, there are significant obstacles that need to be addressed in terms of fulfilling hardware requirements for its implementation.