Time Series

A projected nonlinear state-space model for forecasting time series signals

MEASURE OF DEPENDENCE FOR FINANCIAL TIME-SERIES

TS-Diffusion: Generating Highly Complex Time Series with Diffusion Models

Is Channel Independent strategy optimal for Time Series Forecasting?

Unraveling the ‘Anomaly’ in Time Series Anomaly Detection: A Self-supervised Tri-domain Solution

Multilayer Quantile Graph for Multivariate Time Series Analysis and Dimensionality Reduction

SOLETE Dataset

SWaT WADI EPIC Dataset

Forecasting Methods and Practice Books Online

torchscale_demo

Time series distances

asymmetric loss

MLP-based

DLinear	STID	NHits	Nbeatsx	Nbeats
pdf	pdf	pdf	pdf	pdf
code	code	code	code	code

RNN-based

LSTNet	DeepAR	SRU
pdf	pdf	pdf
code	code	code

CNN-based

SCINet	OmniScale	TCN	mWDN
pdf	pdf	pdf	pdf
code	code	code	code

Transformer-based

PatchTST	ETSformer	FEDformer	Stationary	Scaleformer	Pyraformer	Preformer
pdf	pdf	pdf	pdf	pdf	pdf	pdf
code	code	code	code	code	code	code

TFT	TCCT	Autoformer	Informer	LogTrans	Reformer
pdf	pdf	pdf	pdf	pdf	pdf
code	code	code	code	code	code

wind speed forecast

Multistep short-term wind speed prediction using nonlinear auto-regressive neural network with exogenous variable selection

code

Change-point detection

Change-point detection in wind turbine SCADA data for robust condition monitoring with normal behaviour models

code

Loss function

• TILDE-Q: A Transformation Invariant Loss Function for Time-Series Forecasting
• A Comprehensive Survey of Regression Based Loss Functions for Time Series Forecasting
• Deep Time Series Forecasting with Shape and Temporal Criteria
• Shape and Time Distortion Loss for Training Deep Time Series Forecasting Models

Tricks

• Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift

• ST-Norm: Spatial and Temporal Normalization for Multi-variate Time Series Forecasting

• Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests (tsfresh – A Python package)

• Forecasting at Scale

• NeuralProphet: Explainable Forecasting at Scale

• PyDMD: Python Dynamic Mode Decomposition

• DON’T OVERFIT THE HISTORY -RECURSIVE TIME SERIES DATA AUGMENTATION

• Learning Fast and Slow for Online Time Series Forecasting fsnet

• A Hybrid System Based on Dynamic Selection for Time Series Forecasting code

• Improving forecast stability using deep learning code

loss function (RMSSE for forecast error and RMSSC for forecast instability) instead of three different loss functions (MAPE, sMAPE, and MASE)

Anomaly Detection

Anomaly Detection/Outlier Detection: Why do we need to normalise/standardise our dataset?

Rethinking Graph Neural Networks for Anomaly Detection

Code https://github.com/squareRoot3/Rethinking-Anomaly-Detection

Deep Variational Graph Convolutional Recurrent Network for Multivariate Time Series Anomaly Detection

Code https://github.com/SigmaLab01/DVGCRN

Latent Outlier Exposure for Anomaly Detection with Contaminated Data

Code https://github.com/boschresearch/LatentOE-AD

Time Series Anomaly Detection for Cyber-physical Systems via Neural System Identification and Bayesian Filtering

Code https://github.com/NSIBF/NSIBF

CNKI

THU Number of articles 14893

PKU Number of articles 1548

SJTU Number of articles 65746

ZJU Number of articles 94398

WHU Number of articles 26703

NJU Number of articles 33715

FDU Number of articles 35699

USTC Number of articles 23874

HUST Number of articles 89436

RUC Number of articles 666

Strange

Reading articles written by Indians is not as good as reading articles written by Chinese

Hybrid wind speed forecasting using ICEEMDAN and transformer model with novel loss function[https://doi.org/10.1016/j.energy.2022.126383]

(transformer network (TRA) with a novel KMSE loss function (NLF) for the WSF)

A novel loss function of deep learning in wind speed forecasting[https://doi.org/10.1016/j.energy.2021.121808]

(propose a kernel MSE loss function)

There is no citation relationship in the above two articles