【時序模型】《Deep and Confident Prediction for Time Series at Uber》論文閱讀
摘要 經典的時間序列模型常與不確定性估計的概率公式結合使用。然而這類模型很難調優、縮放和添加外生變量。針對近年來長短期記憶網絡（LSTM）的興起。該文提出了一種新的端到端貝葉斯深度度模型，該模型能夠提供時間徐略預測和不確定性估計。 準確的時間序列預測具有挑戰性的原因在於：1.差異較大的時段。2.因爲極端事件預測依賴於許多外部因素，對於Uber，包括天氣、城市人口增長或營銷變化。這些外部變量很難納入
相關文章
- 1. 論文 Convolutional Networks for Images, Speech, and Time-Series
- 2. EA-LSTM: Evolutionary attention-based LSTM for time series prediction
- 3. A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction閱讀筆記
- 4. Recurrent neural networks for multivariate time series with missing values文獻閱讀
- 5. 文獻閱讀 - Searching and Mining Trillions of Time Series Subsequences under Dynamic Time Warping
- 6. （轉）LSTM NEURAL NETWORK FOR TIME SERIES PREDICTION
- 7. CVPR 2017 Deep representation learning for human motion prediction and classification 論文解讀
- 8. 論文閱讀：Time Series Shapelets: A New Primitive for Data Mining
- 9. 【論文閱讀】 Real-time Personalization using Embeddings for Search Ranking at Airbnb
- 10. 論文閱讀 | Multi-Cast Attention Networks for Retrieval-based Question Answering and Response Prediction
- 更多相關文章...
- • RSS 閱讀器 - RSS 教程
- • ASP.NET MVC - 模型 - ASP.NET 教程
- • 委託模式
- • Kotlin學習（二）基本類型
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
最新文章
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. 論文 Convolutional Networks for Images, Speech, and Time-Series
- 2. EA-LSTM: Evolutionary attention-based LSTM for time series prediction
- 3. A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction閱讀筆記
- 4. Recurrent neural networks for multivariate time series with missing values文獻閱讀
- 5. 文獻閱讀 - Searching and Mining Trillions of Time Series Subsequences under Dynamic Time Warping
- 6. （轉）LSTM NEURAL NETWORK FOR TIME SERIES PREDICTION
- 7. CVPR 2017 Deep representation learning for human motion prediction and classification 論文解讀
- 8. 論文閱讀：Time Series Shapelets: A New Primitive for Data Mining
- 9. 【論文閱讀】 Real-time Personalization using Embeddings for Search Ranking at Airbnb
- 10. 論文閱讀 | Multi-Cast Attention Networks for Retrieval-based Question Answering and Response Prediction