時(shí)序資料匯總：模型和常見庫對比

Part1 領(lǐng)域介紹

Time series is a series of data points indexed in time order.

時(shí)間序列分析具體包括的任務(wù)：

• 檢索Indexing (query by content): given a time series and some similarity measure, find the nearest matching time series.
• 聚類Clustering: find groups (clusters) of similar time series.
• 分類Classification: assign a time series to a predefined class.
• 分割Segmentation (Summarization): create an accurate approximation of a time series by reducing its dimensionality while retaining its essential features.
• 預(yù)測Forecasting (Prediction): given a time series dataset up to a given time tn, forecast the next values.
• 異常檢測Anomaly Detection: find abnormal data points or subsequences.
• 因果分析Rules Discovery: find the rules that may govern associations between sets of time series or subsequences

推薦教材

• Forecasting: Principles and Practice，第三版（英文），第二版（中文）

推薦公開課

• Intel 時(shí)間序列分析：講授時(shí)間序列分析，以及用于預(yù)測、處理和識(shí)別順序數(shù)據(jù)的方法。
• 時(shí)間序列和平穩(wěn)數(shù)據(jù)簡介
• 數(shù)據(jù)平滑化、自相關(guān)性和自回歸積分滑動(dòng)平均 (ARIMA) 模型等應(yīng)用
• 高級(jí)時(shí)間序列概念，如卡爾曼濾波器 (Kalman Filter) 和傅里葉變換 (Fourier Transformation)
• 用于時(shí)間序列分析的深度學(xué)習(xí)架構(gòu)和方法

Part2 時(shí)序Python庫

• Kats，推薦指數(shù)：??
• 主頁：https://facebookresearch.github.io/Kats/
• Github：https://github.com/facebookresearch/Kats
• darts，推薦指數(shù)：??
• 介紹：a Python library for easy manipulation and forecasting of time series. It contains a variety of models, from classics such as ARIMA to deep neural networks.
• 主頁：https://unit8co.github.io/darts/
• Github：https://github.com/unit8co/darts
• GluonTS，推薦指數(shù)：????
• 主頁：https://ts.gluon.ai/index.html
• Github：https://github.com/awslabs/gluon-ts/
• NeuralProphet，推薦指數(shù)：????
• 主頁：https://neuralprophet.com/
• Github：https://github.com/ourownstory/neural_prophet
• arch
• 介紹：Autoregressive Conditional Heteroskedasticity (ARCH) and other tools for financial econometrics, written in Python.
• 主頁：https://arch.readthedocs.io/en/latest/
• Github：https://github.com/bashtage/arch
• AtsPy
• 介紹：Automated Time Series Models in Python
• Github：https://github.com/firmai/atspy
• banpei
• 介紹：Anomaly detection library based on singular spectrum transformation
• Github：https://github.com/tsurubee/banpei
• cesium
• 介紹：end-to-end machine learning platform for time-series, from calculation of features to model-building to predictions.
• 主頁：https://cesium-ml.org/
• Github：https://github.com/cesium-ml/cesium
• pyfbad
• Github：https://github.com/Teknasyon-Teknoloji/pyfbad

更多的模型介紹可以查閱論文[arxiv 2021]A systematic review of Python packages for time series analysis.

Part3 相關(guān)模型

Time Series Forecasting

Time Series Classification

• LSTM FCN，LSTM Fully Convolutional Networks for Time Series Classification

Anomaly Detection

• [AAAI 2022] Towards a Rigorous Evaluation of Time-series Anomaly Detection

Time Series Representation

• [AAAI 2022] TS2Vec: Towards Universal Representation of Time Series

Data Augmentation

• [IJCAI 2021] Time Series Data Augmentation for Deep Learning: A Survey
• [arxiv 2020] An empirical survey of data augmentation for time series classification with neural networks

Part4 時(shí)序數(shù)據(jù)集

• UCR Time Series Classification Archive
• UEA & UCR Time Series Classification Repository