메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(Korea University) (Korea University)
저널정보
한국데이터정보과학회 한국데이터정보과학회지 한국데이터정보과학회지 제36권 제3호
오류 신고하기

피인용 0

검색

    초록·키워드

    This study introduces two categories of time series models: Transformer-based models (Transformer, Reformer, Informer, and Autoformer) and LLM-based models (Prompt-cast, LLMTime, and Time-LLM). They have been actively researched as solutions to address the long-term dependency issues and computational costs arising from the lack of data parallelization in traditional time series models. These models are applied to predict login frequency on platforms used by universities for academic credit exchange, enabling a comparative analysis of their predictive performance. The results indicate that Transformer-based models generally achieve superior forecasting performance compared to LLM-based models; however, they also require significantly longer computational time. Specifically, the LLM-based models were implemented using GPT-3.5 and GPT-4o.

    최근 본 자료 전체보기

      UCI(KEPA) : I410-151-25-02-093148781