International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 13 Issue: 06 | Jun 2026
p-ISSN: 2395-0072
www.irjet.net
Crypto Predict: AI-Powered Crypto currency Price Prediction & Sentiment Analysis 1Samruddhi Gahininath Bargaje, 2Bhanage Vrushali Machhindra, 3Rachana Bhimraj Mandlik, 4Sakshi Sadashiv Mane
----------------------------------------------------------------------------***-------------------------------------------------------------------------sentiment [3][4]. These factors make accurate prediction Abstract-The rapid growth of cryptocurrency markets of cryptocurrency prices a highly challenging task. has introduced significant volatility, making accurate Traditional financial forecasting models, which rely price prediction a challenging task. This paper presents heavily on statistical assumptions and linear Crypto Predict, an AI-powered system that integrates relationships, often fail to capture the nonlinear and machine learning and natural language processing dynamic nature of cryptocurrency markets [5]. As a techniques to forecast crypto currency prices and analyse result, machine learning (ML) and artificial intelligence market sentiment. The proposed system utilizes historical (AI) techniques have gained significant attention due to price data along with technical indicators such as Simple their ability to model complex patterns, learn from large Moving Averages (SMA) and Relative Strength Index (RSI) datasets, and adapt to changing market conditions [6][7]. to predict short-term price movements using an Various ML approaches, including regression models, autoregressive linear regression model. Additionally, decision trees, support vector machines, and deep sentiment analysis is performed on social media data learning architectures, have been applied to predict using TextBlob to capture market psychology and investor cryptocurrency price movements with improved behavior. The system combines quantitative price accuracy [8][9]. predictions with qualitative sentiment insights to generate intelligent trading signals such as Buy, Sell, or Hold. A fullIn addition to quantitative price data, qualitative factors stack architecture is implemented using Flask for backend such as public sentiment and investor behavior play a processing, React for frontend visualization, and MongoDB crucial role in influencing cryptocurrency markets. Social for data storage. Real-time cryptocurrency data is fetched media platforms, online forums, and news articles often via API integration to ensure up-to-date predictions. drive market trends by shaping investor perceptions and Experimental results demonstrate high prediction decisions [10][11]. Consequently, sentiment analysis accuracy and effective sentiment classification, improving using Natural Language Processing (NLP) techniques has decision-making reliability. The proposed approach emerged as an essential tool for understanding market highlights the importance of hybrid models that integrate psychology. Methods such as TextBlob, VADER, and financial indicators with behavioural analysis for transformer-based models enable the extraction of enhanced cryptocurrency forecasting and trading sentiment polarity and emotional context from textual strategies. data [12][13]. Keywords-Crypto currency Prediction, Machine Recent research has focused on hybrid approaches that Learning, Sentiment Analysis, Linear Regression, combine machine learning-based price prediction with Natural Language Processing, Financial Forecasting, sentiment analysis to improve forecasting performance Trading Signal Generation [14]. These integrated systems leverage both technical indicators (e.g., moving averages, RSI, MACD) and I. INTRODUCTION sentiment signals to provide a more comprehensive understanding of market dynamics [15]. However, many Cryptocurrency markets have experienced exponential existing solutions lack real time data integration, growth over the past decade, emerging as a significant scalability, and unified frameworks for generating component of the global financial ecosystem. Unlike actionable trading decisions. traditional financial systems, cryptocurrencies operate on decentralized blockchain technology, eliminating the To address these limitations, this paper proposes Crypto need for centralized authorities and enabling peer-toPredict, an AI-powered hybrid system that combines peer transactions [1][2]. This decentralization, combined autoregressive machine learning models with sentiment with the potential for high returns, has attracted analysis for cryptocurrency price prediction. The system investors, researchers, and institutions worldwide. incorporates historical price data, real-time API inputs, However, despite their advantages, crypto currency and social media sentiment to generate accurate markets are characterized by extreme volatility, rapid forecasts and intelligent trading signals such as Buy, Sell, price fluctuations, and susceptibility to external and Hold. By integrating quantitative and qualitative influences such as regulatory changes, technological analysis within a full stack architecture, the proposed developments, macroeconomic trends, and investor
© 2026, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 215