Presentation Proposal

[Virtual Presenter] Comparison of Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) Algorithms for Bitcoin Price Prediction 11321007 - Angelica Theresia Manurung 11321021 - Luana Breka Banjarnahor 11321051 - Horas Marolop Amsal Siregar Dosen Pembimbing : Tegar Arifin Prasetyo, S.Si., M.Si Institut Teknologi Del D3 Teknologi Informasi Fakultas Vokasi.

[Audio] Literature Review Research Question Purpose Scope OUTLINES Background 1 2 3 4 5 Research Methodology 6.

[Audio] BACKGROUND Bitcoin is famous for its high fluctuations where the price can rise or fall dramatically quickly. For this reason, investors must understand the relationship between risk and return of an investment before deciding to invest. To anticipate the unstable price of Bitcoin, a prediction or forecasting of the price of Bitcoin in the future is carried out. So that investors do not need to be wary of any price changes and can make a profit..

[Audio] LITERATURE REVIEW ❏ Chih Hung, Wu et al (2018) focused on the problem of bitcoin currency prediction with the help of LSTM algorithm and ARIMA model. The problem includes the development of a system to analyze the statistics of the Bitcoin currency price. The prediction results of the two models each calculate the MAPE evaluation indicator of 11.86% with ARIMA and MAPE 1.40% with conventional (LSTM). The resulting average error rate (MAPE) shows that the use of the LSTM algorithm is very effective, because the number is below 10%, indicating the research results are very accurate. ❏ Hasan, H et al (2022) focused on the problem of predicting cryptocurrency prices based on user opinions from social media using the CNN algorithm. The results of this study show that the proposed method achieves high accuracy of about 98.75% in price prediction, but this study is not very efficient in predicting prices for the next 7 days. This shows that CNN is more suitable for short-term prediction than long-term prediction..

[Audio] LITERATURE REVIEW ❏ I. Indera, N et al (2018) focused on the problem of nonlinear autoregressive bitcoin price prediction with exogenous input (NARX) using PSO optimized parameters and Moving Average technical indicators. The results of Bitcoin price prediction using MPL-based NARX prediction show minimal coefficients within the 95% confidence limit. Therefore, the NARX model is accepted as a valid model for predicting bitcoin price. ❏ Srivastava, V et al (2023) focused on Cryptocurrency price prediction problem using PSO enhanced with Extreme Gradient algorithm. The experimental results of the model performance analysis showed low percentage metric values of 0.000407408, 0.030922822, 0.037967598 (MSE), 0.000399667, 0.071487825, 0.006800319 (MAE), 0.002018434, 0.017584886, 0.019485276 (RMSE) for Dogecoin, Bitcoin, and Ethereum cryptocurrencies. This shows the accurate predictive ability of the estimated closing prices of the cryptocurrencies..

[Audio] RESEARCH QUESTION 1. How to build a website in predicting the price of Bitcoin using the Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) algorithms? 2. How does the Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) algorithms work in predicting the price of Bitcoin? 3. What is the accuracy level of Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) in predicting Bitcoin price?.

[Audio] 02 03 01 PURPOSE Build a web-based application that is used to predict the price of Bitcoin using the Particle Swarm Optimization (PSO) algorithm and the Long Short Term Memory (LSTM) algorithm. Analyze how the Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) algorithms work in predicting Bitcoin prices. Knowing the final results of the accuracy level of the Particle Swarm Optimization (PSO) algorithm and the Long Short Term Memory (LSTM) algorithm in predicting Bitcoin prices..

[Audio] SCOPE 1 Implement deep learning algorithms, namely the Particle Swarm Optimization (PSO) and Long Short Term Memory (LSTM) algorithms in predicting Bitcoin prices. 2 The model built will use the CoinMarketCap dataset based on data from January 2021 to December 2023. 3 The model was built using the Django framework with the python language..

[Audio] RESEARCH METHODOLOGY CRISP-DM Business understanding Data understanding Data Preprocessing Modelling Evaluation Deployment Current understanding of Bitcoin business Analyzing Bitcoin price in terms of potential or characteristics. Perform testing with steps from data preprocessing data The use of algorithmic models in predicting Bitcoin price Define an evaluation matrix to evaluate the performance of the model Perform the designed implementation by using PSO and LSTM algorithms.

[Audio] CONCLUSION From previous research, the use of PSO and LSTM algorithms provides good and consistent Bitcoin price prediction results. For this reason, in this study, a comparison of the PSO and LSTM algorithms is carried out to determine which algorithm is suitable for predicting Bitcoin prices with the accuracy value of the algorithm's prediction results..

[Audio] THANK YOU GROUP 10. THANK YOU. GROUP 10.