AI/Abnormal Detection8 [transformer] PatchTST https://youtu.be/VrqkXrwpohw?si=1u-Vp-IOCTEj_vhC TSF(TimeSeries Forecasting) Task 내 딥러닝 모델 한계점 Vanila-Transformer Vanila-Transformer는 long sequence에서 sematic correlation을 추출하는데 효과적이나, self-attention구조상 permutaion-invariant 특성을 가지기 때문에 temporal loss가 발생 permutaion-invariant 특성은 input의 순서가 바뀌어도 output이 동일한 것을 의미. 시계열 데이터는 데이터 순서가 중요한 정보인데, self-attention을 통해 input과 output 의미가 퇴색 .. 2025. 2. 17. [Transformer] Time Series Classification model based on Transformer * https://youtu.be/1ggV-0Y0rkE?si=Q-tCs_q2Xf4VIktU * https://youtu.be/D8BXliGyUSM?si=PbqzIkmE8jlwhXwc * https://www.linkedin.com/pulse/time-series-classification-model-based-transformer-gokmen Time Series Classification model based on TransformerRecent studies show that transformer models numerous contribution on several tasks such as classification, forecasting and segmentation. This article .. 2024. 12. 17. [keras] multivariate time series - CNN Conv2D- LSTM Keras time series prediction with CNN+LSTM model and TimeDistributed layer wrapper Keras time series prediction with CNN+LSTM model and TimeDistributed layer wrapperI have several data files of human activity recognition data consisting of time-ordered rows of recorded raw samples. Each row has 8 columns of EMG sensor data and 1 corresponding column of target ...stackoverflow.com CNN + LSTM 아키텍.. 2024. 12. 12. Time Series Clustering - K-Means + Dynamic Time Warping (비지도 기반 시계열 데이터 군집 시각화) http://dmqm.korea.ac.kr/activity/seminar/245 고려대학교 DMQA 연구실고려대학교 산업경영공학부 데이터마이닝 및 품질애널리틱스 연구실dmqa.korea.ac.kr 2024. 12. 4. 이전 1 2 다음
