상세정보
학술대회
A Study on Online ARIMA Algorithms Applying Various Gradient Descent Optimization Algorithms for Time Series Prediction
- 발행일
- 202110
- 출처
- International Conference on Information and Communication Technology Convergence (ICTC) 2021, pp.1104-1106
- 협약과제
- 21IK1100, 재생에너지/분산전원(ESS/EV)/에너지수요 BTM 단위 모니터링 및 AI를 활용한 마이크로 그리드 친환경 에너지 예측/예보 기술개발, 이병탁
- 초록
- The prediction of time series data can be used in various fields, and many research for this is required. This paper has researched about applying various optimizers for online ARIMA model, which is used to make a prediction on time series data. For this work, algorithms for gradient descent like Adam, was used for updating online ARIMA model's weight for predicting the values. Our approach to finding a gradient descent algorithm that can obtain higher performance in time series prediction, the performance of the online ARIMA-based model was evaluated with the photovoltaic power generation data and gold price data. As a result, it was confirmed that there is a gradient descent-based optimization algorithm that improves the prediction performance for each time-series data.
- KSP 제안 키워드
- ARIMA model, Gradient descent algorithm, Gradient descent optimization, Higher performance, Optimization algorithm, Time series data, gold price, photovoltaic power generation, prediction performance, time series prediction