- 學位論文
基於Ridge回歸模型預測非同質化代幣之價格
Price Prediction of Non-Fungible Token Using Ridge Regression
摘要
隨著Web3.0與區塊鏈技術的興起，有了虛擬貨幣的誕生，也因此掀起炒幣熱潮，由區塊鏈的底層技術延伸出的非同質化代幣(NFT)，其獨一無二且不能被分割或替換的特性，具有極高的稀缺性，帶動了數位藝術市場，使NFT成為投資市場的焦點。對於NFT而言，衡量NFT之藝術價值是一件困難的事情，因為市場潮流總會隨著時間變化，人們的喜好也會隨之改變，NFT的定價確實會受到多項重要因素的影響。 本研究主要是對於非同質化代幣(NFT)進行價格預測，針對SuperRare NFT平台的藝術圖像進行脊迴歸(Ridge)模型分析，將資訊熵(entropy)與機器學習方法相結合，對於稀缺性將提出衡量方法，藉此發展NFT定價模式，其能有效預測NFT之價值。 研究結果顯示，出價對於NFT的最終售價有正向影響，是一項重要控制變數。若加入稀缺性、藝術鑑賞及複雜度等作為特徵變數，預測能力皆會提高，說明NFT的定價並非只靠某一種因素而決定，出價、稀缺性和藝術鑑賞等都是能夠影響NFT售價的重要因素。
並列摘要
With the advent of Web3.0 and the emergence of blockchain technology, the birth of virtual currencies has sparked a fervor in the realm of speculative trading. Non-fungible tokens (NFTs), derived from the foundational underpinnings of blockchain, possess inherent qualities that are unparalleled and indivisible, impervious to division or substitution. Their remarkable scarcity has propelled the digital art market, catapulting NFTs into the very focal point of investment. As for NFTs, measuring their artistic worth proves to be a formidable undertaking, given that market trends inevitably fluctuate over time, accompanied by the ever-evolving tastes and preferences of individuals. Indeed, the pricing of NFTs is undeniably influenced by a multitude of pivotal factors.This study primarily delves into price prediction for NFTs. It conducts a ridge regression model analysis specifically focused on the artistic imagery found on the SuperRare NFT platform. By amalgamating the concept of information entropy with machine learning methodologies, a methodology to gauge scarcity shall be proposed, thereby forging an NFT pricing model that effectively prognosticates their value.The research findings unveil a positive correlation between bidding and the ultimate selling price of NFTs, deeming bidding to be a relatively crucial predictor of the price. It stands as a significant control variable. Incorporating variables such as scarcity, artistic appreciation, and complexity enhances the predictive capabilities, elucidating that the pricing of NFTs is not solely dictated by any single factor. Bidding, scarcity, and artistic appreciation collectively stand as vital constituents capable of influencing the selling price of NFTs.