1. Related
2. Challenges
3. Tutorials
4. Books
5. Papers
6. Conference
7. Scholar
8. Competitions
9. STOA
10. Datasets
11. Tools

• Information Theory

• Signal Processing/Digital Signal Processing

  time frequency analysis, fourier analysis, wavelets,...

• Audio Content Analysis

  fundamentals of sound and time-frequency representations, periodicity detection, novelty detection, sound classification, ...

  http://www.nyu.edu/classes/bello/Teaching.html

• Dynamical Systems Theory

• Complex Seasonal Patterns

  Forecasting time series with complex seasonal patterns using exponential smoothing, Alysha M De Livera, Rob J Hyndman and Ralph D Snyder

  http://businessforecastblog.com/analyzing-complex-seasonal-patterns/

• Mixed Frequency

  Mixed data sampling (MIDAS) models

  Forecasting with mixed frequencies

• Irregularly Sampled (Unevenly Spaced, Sparse)

  https://en.wikipedia.org/wiki/Unevenly_spaced_time_series

  Comparison of correlation analysis techniques for irregularly sampled time series

  http://www.eckner.com/research.html

  AWarp: Warping Distance for Sparse Time Series site

• Hierarchical

  http://robjhyndman.com/papers/hierarchical/

  http://www.forecastpro.com/Trends/forecasting101January2009.html

• Highly Frequency

  https://www.zhihu.com/question/26464548

1. Architecture
2. Property
3. Feature
4. Visualization
5. Model
6. Strategy
7. Topic
8. Application
9. Q&A

timeseries structure design

summary from the competitions experience.

• Random Walk

  https://www.kaggle.com/thebrownviking20/everything-you-can-do-with-a-time-series/

• Stationary

  mean stationary, variance stationary

• The Ergodic Theorem

  the time series chapter in this book

• The Takens's theorem

  https://en.wikipedia.org/wiki/Takens%27_theorem

• Attractor

  Attractor reconstruction - Scholarpedia

  Chaotic Attractor Reconstruction - node99

• Motif

  Finding Motifs in Time Series

  Exact discovery of time series motifs site

  Detecting time series motifs under uniform scaling paper site

• Periodicity

  periodicity detection/estimation:

  Multi-step approach to find periods of time-series data site

  detecting multiple periodicity in time series site

• Embedding Dimension

  The false nearest neighbors algorithm: An overview

• SAX(Symbolic Aggregate approXimation)

  SAX is the first symbolic representation for time series that allows for dimensionality reduction and indexing with a lower-bounding distance measure. In classic data mining tasks such as clustering, classification, index, etc., SAX is as good as well-known representations such as Discrete Wavelet Transform (DWT) and Discrete Fourier Transform (DFT), while requiring less storage space. http://www.cs.ucr.edu/~eamonn/SAX.htm

  Experiencing SAX: a novel symbolic representation of time series

  HOT SAX: Efficiently Finding the Most Unusual Time

• Topological

  Topological Time Series Analysis:

  Geometry of sliding window embeddings

  Persistent Homology of Sliding Window Point Clouds

  https://www.joperea.com/

• data transformation

  http://people.duke.edu/~rnau/whatuse.htm

• time series features/structure:

  characteristics

  features of time series

  Finding Repeated Structure in Time Series Algorithms and Applications site | site

  Slow feature analysis: Unsupervised learning of invariances

• feature extraction/transformation

  time-based feature construction part1 part2 part3

  summary from the competitions experience.

  Automatic Identification of Time Series Features for Rule-Based Forecasting

  Distributed and parallel time series feature extraction for industrial big data applications

• feature selection

  Feature selection for time series prediction – A combined filter and wrapper approach for neural networks

• Recurrence Plots

• Kalman filter

  https://en.wikipedia.org/wiki/Kalman_filter
  http://www.cs.unc.edu/~welch/kalman/ http://www.cl.cam.ac.uk/~rmf25/papers/Understanding%20the%20Basis%20of%20the%20Kalman%20Filter.pdf http://www.bzarg.com/p/how-a-kalman-filter-works-in-pictures/ http://blog.csdn.net/xiahouzuoxin/article/details/39582483

• Fourier transform

  An Interactive Guide To The Fourier Transform

  The Fast Fourier Transform - Math ∩ Programming

  Understanding the FFT Algorithm

  http://news.mit.edu/2009/explained-fourier http://news.mit.edu/2012/faster-fourier-transforms-0118

  我所理解的快速傅里叶变换（FFT）

• Dynamic time warping

  A Bibliography of Dynamic Time Warping site

  Extracting Optimal Performance from Dynamic Time Warping site

  Everything you know about Dynamic Time Warping is Wrong

  Searching and Mining Trillions of Time Series Subsequences under Dynamic Time Warping

  https://gradientflow.com/2012/10/28/mining-time-series-with-trillions-of-points-dynamic-time-warping-at-scale/

  https://blog.acolyer.org/2016/05/11/searching-and-mining-trillions-of-time-series-subsequences-under-dynamic-time-warping/

• Simple model (Such as mean, moving averages,...)

  http://people.duke.edu/~rnau/whatuse.htm

  The mean (constant, intercept-only) model for forecasting

  Review of basic statistics and the simplest forecasting model: the sample mean

  https://people.duke.edu/~rnau/411mean.htm

  Simple versus complex forecasting: The evidence

  Benchmarks for forecasting - Hyndsight

  Moving averages paper

  Theta:

  The theta model - a decomposition approach to forecasting

  Unmasking the Theta method

  Case:

  Rossmann sales forecasting(mean and weights) https://github.com/Dyakonov/notebooks

  数据竞赛思路分享：机场客流量的时空分布预测 - ZJun Thinking - 博客频道 - CSDN

  如何在第一次天池比赛中进入Top 5%（一） - 知乎专栏

• X13-SEATS-ARIMA

• TBATS

• ARIMAX

• ARX-ARMAX

• ARDL(Auto Regressive Distributed Lag) blog

• Dynamic Regression Models

• ...

• SVM

  Time series prediction using support vector machines: a survey

  Support vector machines experts for time series forecasting

  Predicting time series with support vector machines

  Application of support vector machines in financial time series forecasting

• Boosting

  A gradient boosting approach to the Kaggle load forecasting competition

• NARX

  Long-Term Time Series Prediction with the NARX Network An Empirical Evaluation

• Bayesian neural network (BNN)

  Deep and Confident Prediction for Time Series at Uber arxiv

  https://eng.uber.com/neural-networks-uncertainty-estimation/

• Neural Network

  http://www.neural-forecasting.com/tutorials.htm

  Designing a neural network for forecasting financial time series

  Neural network forecasting for seasonal and trend time series

  RNN:

  Engineering Extreme Event Forecasting at Uber with Recurrent Neural Networks

  LSTM:

  Long Short Term Memory Networks for Anomaly Detection in Time Series

• Hybrid

  Time series forecasting using a hybrid ARIMA and neural network model

  An artificial neural network (p,d,q) model for timeseries forecasting

• Forecast Combination

  Timmermann, Allan, 2006. "Forecast Combinations," Handbook of Economic Forecasting, Elsevier. ppt

  A simple explanation of the forecast combination puzzle

  Combining time series models for forecasting

  Optimal combination forecasts for hierarchical time series

• Machine Learning Strategies for Time Series Prediction slideshare

• Machine learning strategies for multi-step-ahead time series forecasting PhD thesis

• Training Strategies for Time Series: Learning for Prediction, Filtering, and Reinforcement Learning thesis

• MASE

  Another look at measures of forecast accuracy

• Cross-Validation

  Time Series Nested Cross-Validation

  On the use of cross-validation for time series predictor evaluation

  A note on the validity of cross-validation for evaluating autoregressive time series prediction

  Approximate leave-future-out cross-validation for Bayesian time series models

• Prediction Intervals

  The difference between prediction intervals and confidence intervals - Hyndsight

  Why time series forecasts prediction intervals aren't as good as we'd hope

  Better prediction intervals for time series forecasts

  Prediction intervals for ensemble time series forecasts

• Spectral analysis

  Spectral analysis is carried out to describe how variation in a time series may be accounted for by cyclic components. This may also be referred to as "Frequency Domain". With this an estimate of the spectrum over a range of frequencies can be obtained and periodic components in a noisy environment can be separated out.

• Intervention Analysis

  Time Series Intervention Analysis (or Interrupted Time Series Analysis) can explain if there is a certain event that occurs that changes a time series. This technique is used a lot of the time in planned experimental analysis.

  The basic question is "Has an event had an impact on a time series?"

• Calendar effects

  http://calendar-effects.behaviouralfinance.net/

  Special days, Holidays,...

  Public_holidays_in_China

• Causality

  Convergent Cross Mapping

  https://github.com/NickC1/skCCM

• Similarity

  https://en.wikipedia.org/wiki/Cross-correlation

  Detect correlation between multiple time series - Anomaly

  Similarity Search on Time Series Data: Past, Present and Future site

  Mueen's Algorithm for Similarity Search site

  Querying and Mining of Time Series Data Experimental Comparison of Representations and Distance Measures

  Experimental Comparison of Representation Methods and Distance Measures for Time Series Data

• Cluster

  Clustering of time series data—a survey

  Clustering of Time Series Subsequences is Meaningless Implications for Previous and Future Research

  Time-series clustering – A decade review

  Dynamic Time Warping Clustering - Cross Validated

  k-Shape-Efficient and Accurate Clustering of Time Series, SIGMOD 2015, site

• Classification

  Highly comparative feature-based time-series classification

• Anomaly Detection

  Outlier Detection for Temporal Data: A Survey

  Outlier Detection for Temporal Data, Manish Gupta, Microsoft India and IIIT, Jing Gao, SUNY, Buffalo, Charu Aggarwal, IBM TJ Watson, Jiawei Han, UIUC - book

  https://github.com/rob-med/awesome-TS-anomaly-detection

  http://stats.stackexchange.com/questions/137094/algorithms-for-time-series-anomaly-detection

  Time Series Anomaly Detection Algorithms – Stats and Bots

  https://github.com/twitter/AnomalyDetection

  https://github.com/twitter/BreakoutDetection

  Anomaly Detection for Airbnb's Payment Platform - Airbnb Engineering

  https://anomaly.io/about/index.html

  Time-series novelty detection using one-class support vector machines

• time space

  大佬用四句五个字来概括了这个领域的整体问题： 空间不变性 空间可变性 时间不变性 时间可变性

• Electricity

  Electricity price forecasting: A review of the state-of-the-art with a look into the future

  A neural network based several-hour-ahead electric load forecasting using similar days approach

  Modelling and forecasting daily electricity load curves: a hybrid approach

  Combined modeling for electric load forecasting with adaptive particle swarm optimization

  Triple seasonal methods for short-term electricity demand forecasting

  Short-term forecasting of anomalous load using rule-based triple seasonal methods

  Rule-based autoregressive moving average models for forecasting load on special days: A case study for France

• Business

  Business Forecasting Practical Problems and Solutions, Edited by Michael Gilliland, Len Tashman, Udo Sglavo

• http://www.forecastingprinciples.com/index.php/faq

• Answers to Frequently Asked Questions

• Pitfalls in time series analysis - Cross Validated

• Is it unusual for the MEAN to outperform ARIMA? site

• How to know if a time series is stationary or non-stationary - Cross Validated

• When to log transform a time series before fitting an ARIMA model site

• Don’t Put Lagged Dependent Variables in Mixed Models site

• Best method for short time-series site

• Estimating same model over multiple time series site

• correlating volume time series site

• correlation between two time series site

• Is it possible to do time-series clustering based on curve shape? site

• features for time series classification site

• Modelling longitudinal data where the effect of time varies in functional form between individuals site

• Why can't we use top-down methods in forecasting grouped time series? site

• Proper way of using recurrent neural network for time series analysis site

• Does the DTW method consider the similarity in scale and time delay of two time series ?

• simple algorithm for online outlier detection of a generic time series site

• outliers spotting in time series analysis should i pre-process data or not site

• how to adjusting chinese new year effects site

• how to treat holidays when working with time series data site

• Using k-fold cross-validation for time-series model selection - Cross Validated

• Time Series Nested Cross-Validation – Towards Data Science

• Interpretation of mean absolute scaled error (MASE) - Cross Validated

• 25 Years of Time Series Forecasting, Jan G De Gooijer, Rob J Hyndman
• A review on time series data mining, Tak Chung Fu
• A Survey on Nonparametric Time Series Analysis, Siegfried Heiler
• Segmenting Time Series: A Survey and Novel Approach

• https://github.com/bighuang624/Time-Series-Papers

• Introdcution

  https://en.wikipedia.org/wiki/Forecasting

  Statistical forecasting: notes on regression and time series analysis site ⭐⭐⭐⭐⭐

  An Introductory Study on Time Series Modeling and Forecasting

  Highly comparative time-series analysis the empirical structure of time series and their methods

• Time Series Analysis, James D. Hamilton, Princeton University Press, 1994

• Time Series Analysis Forecasting and Control (5th Edition), George E. P. Box

• Principles of Forecasting: A Handbook for Researchers and Practitioners, Editors: Armstrong, J.S. (Ed.)

• Forecasting: Principles and Practice (2ed), Rob J Hyndman and George Athanasopoulos online ⭐⭐⭐⭐⭐

  Notes: the ETS (Error, Trend, Seasonal) framework

• Forecasting with Exponential Smoothing: The State Space Approach, Hyndman, R.J., Koehler, A.B., Ord, J.K., Snyder, R.D. online

• Analysis of Financial Time Series (3ed), Ruey S. Tsay site

• The Elements of Financial Econometrics site

• Nonlinear Time Series Nonparametric and Parametric Methods site

• Nonlinear Time Series Analysis

• https://forecasters.org/

• Makridakis

  Pioneered Empirical competition on Forecasting called M, M2 and M3, and paved way for evidence based methods in forecasting

• J. Scott Armstrong site ⭐⭐⭐⭐⭐

  Provides valuable insights in the form of books/articles on Forecasting Practice

  Simple versus complex forecasting: The evidence

  Rule-Based Forecasting: Using Judgment in Time-Series Extrapolation

  Combining Forecasts

  Standards and Practices for Forecasting

• Kesten C. Green site forecasting principles simple forecasting ⭐⭐⭐⭐⭐

  Unifying theory of forecasting: The Golden Rule of Forecasting provides a unifying theory of forecasting. The Rule is to be conservative when forecasting by relying on cumulative knowledge about the situation and about forecasting. Following the Golden Rule guidelines reduces forecast errors by nearly a third, on average, compared to common practice. Superiority of simple forecasting methods: Sophisticatedly simple forecasting methods, which can be understood by decision makers, reduce forecast errors by nearly a quarter, on average, compared to forecasts from complex statistical methods.

  ...

  Golden Rule of Forecasting: Be conservative

  Golden Rule of Forecasting Rearticulated: Forecast Unto Others as You Would Have Them Forecast Unto You

  ...

• Gardner

  Invented Damped Trend exponential smoothing another simple method which works surprisingly well vs. ARIMA

• Eamonn Keogh site tutorials ⭐⭐⭐⭐⭐

  Dynamic time warping

• Rob J Hyndman blog github ⭐⭐⭐⭐⭐

  Some interesting topics on the blog: forecast intervals for aggregates (the aggregate of several time periods), fitting models to short time series, fitting models to long time series, forecasting weekly data, forecasting with daily data, forecasting with long seasonal periods, seasonal periods, rolling forecasts, batch forecasting, facts and fallacies of the AIC, cross-validation, ...

• Tao Hong blog

  Specialize: energy forecasting, electric load forecasting

  PhD thesis: Short Term Electric Load Forecasting

  Course: Electric Load Forecasting I: Fundamentals and Best Practices, Electric Load Forecasting II: Advanced Topics and Case Studies

  Some interesting topics on the blog: forecasting and backcasting,...

• 施行建

  香港中文大学。主要研究的方向是时空序列问题，时间维度为主，并且降水预测的应用。

• 郑宇

  JD的副总裁。主要研究时空序列数据挖掘，空间维度为主，并且用在traffic flow上的应用。主页：http://urban-computing.com/yuzheng

• The M-competitions dataset
• The Tourism Forecasting Competition dataset
• Global Energy Forecasting Competition (GEFCom)
• http://www.neural-forecasting-competition.com/index.htm

• https://www.paperswithcode.com/area/time-series

• UCR Time Series Classification Archive http://www.cs.ucr.edu/~eamonn/time_series_data/

• CompEngine

  A self-organizing database of time-series data

  https://www.comp-engine.org/#!

• R package

  forecast, tsfeatures, thief(Temporal Hierarchical Forecasting), tsDyn, ForecastCombinations, forecastHybrid, opera(Online Prediction by ExpeRt Aggregation)

• awesome_time_series_in_python

• prophet

• tsfresh

• palladium

  Framework for setting up predictive analytics services

• Forecast Pro

• Autobox

• TISEAN - Nonlinear Time Series Analysis

• Timeseries analysis for neuroscience data http://nipy.org/nitime

  https://github.com/nipy/nitime

• Tensorflow

  https://github.com/tgjeon/TensorFlow-Tutorials-for-Time-Series

  https://github.com/mouradmourafiq/tensorflow-lstm-regression

• Keras

  https://github.com/jaungiers/LSTM-Neural-Network-for-Time-Series-Prediction

  https://github.com/cauchyturing/UCR_Time_Series_Classification_Deep_Learning_Baseline

  https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/