KaRuNa: A Blockchain-Based Sentiment Analysis Framework for Fraud Cryptocurrency Schemes

Abstract

The current open cryptocurrency markets pose varied challenges on a prospective investor (PI), such as pseudoanonymity of cryptocurrency transactions, selection criteria for investments in crowdfunding schemes (CF), modus-operandi for these schemes, non-transparency of money generation and distribution among peers, and untraceable scams. PIs are susceptible to monetary losses in the open market due to the aforementioned issues. The fraudsters could be both internal (operator of the scheme) and external (financial institutions (FI), such as banks, money-lenders, and insurance companies). The centrality of trust among stakeholders like PI, CF, and FI is a prime concern. Motivated from these facts, this paper proposes a decentralized framework, KaRuNa, A Blockchain-based Sentiment analysis framework for Fraud Cryptocurrency schemes. KaRuNa operates on public blockchain three phases of trust modeling among stakeholders. In the first phase, transactions are performed on the blockchain that offers trust, auditability, and transparency among stakeholders. In the second phase, sentiment analysis (SA) of cryptocurrencies is proposed based on a novel algorithm of hash addresses to generate classification scores (CS). Parameters like social trends, rise/fall in cryptocurrency price, measured standard deviation, peak and low are selected to fed to proposed novel Long-short term memory (LSTM) classifier to generate recommendations based on CS. An accuracy of 98.99% is achieved using LSTM over generated CS to evaluate risks in the investment. Results demonstrate that KaRuNa achieves more scalability compared to conventional approaches.