Chatbot Informasi Cryptocurrency dan Forecast Cryptocurrency Menggunakan RNN-LSTM

Abstract

Cryptocurrencies have garnered significant attention in the financial market, with rapidly changing and unpredictable price fluctuations. To assist cryptocurrency investors in making informed decisions or obtaining information, many chatbots have been developed as tools to provide insights into cryptocurrency market analysis. However, some chatbots currently rely on simplistic models that struggle to cope with the volatile nature of cryptocurrencies and may not provide real-time data. This research aims to enhance the capabilities of chatbots in delivering information about the cryptocurrency market and forecasting cryptocurrency prices by implementing Long Short-Term Memory (LSTM), an RNN neural network capable of addressing sequential issues and modeling long-term relationships in real-time data. The implementation of LSTM in chatbots is expected to improve the accuracy of cryptocurrency price forecasts, thereby aiding users in making cryptocurrency investment decisions. The research methodology includes accurate cryptocurrency price data collection, data processing, the development of LSTM-based chatbots, and chatbot performance evaluation. The results of this study are expected to enhance the chatbot's ability to provide real-time information about cryptocurrencies to users and offer more accurate price forecasts. Consequently, this research will make a positive contribution to understanding the complex cryptocurrency price movements and assist users in making more intelligent investment decisions in the highly dynamic and volatile cryptocurrency market. Based on this research, the chatbot implemented using the Long Short-Term Memory (LSTM) model succeeded in increasing accuracy in providing information about the crypto market and predicting prices with a success rate of 90% and forecasting cryptocurrency prices had a change percentage of 3,8% .