Time Series Analysis with R
1
Introduction
2
Time series basics
2.1
What is a Time Series
2.1.1
Definition
2.2
Time Series exploration
2.3
Time Series patterns
2.3.1
How to extract the trend, seasonality and error?
2.3.2
How to de-trend a time series ?
2.3.3
How to de-seasonalize a time series in R?
2.4
Stationarity in Time Series
2.4.1
Elementary statistics
2.4.2
Stationarity definition
2.4.3
Stationarity test
2.4.4
Make a time series stationary
2.5
Auto-correlation
2.6
References
3
ARMA Time Series modeling
3.1
Auto-Regressive Time Series model
3.2
Moving Average Time Series Model
3.3
Model selection: AR or MA
3.4
References
4
Time Series Forecasting
4.1
Understanding the context of forecasting
4.2
Data preparation
4.3
Case study
4.4
Performance metrics
4.4.1
Forecast errors
4.4.2
R-squared
4.4.3
Mean Absolute Error (MSE)
4.4.4
Median Absolute Error (MedAE)
4.4.5
Mean Squared Error (MSE)
4.5
Naive methods
4.6
Exponential smoothing
4.6.1
State Space Models
4.6.2
Double seasonal Holt-Winters
4.7
ARIMA/SARIMA models
5
Outlier detection in Time series
5.1
Introduction
5.1.1
Definition
5.1.2
Taxonomy
5.1.3
Types of anomalies in time series
5.1.4
Methodological Approaches
5.2
Statistical-based approaches
5.2.1
Descriptive statistics
5.2.2
Statistical tests
5.2.3
STL decomposition
5.2.4
Generalized ESD Test for Outliers
5.2.5
Extreme Studentized Deviate Technique (ESD)
5.3
Forecasting-based approaches
5.3.1
Moving Average Method
5.3.2
ARMA
5.3.3
Prophet
5.4
Neural Network Based Approaches
5.4.1
Autoencoder
5.5
Clustering Based Approaches
5.5.1
Kmeans
5.5.2
Gaussina Mixture Model (GMM)
5.5.3
DBSCAN
5.6
Proximity Based Approaches
5.6.1
K-Nearest neighbor:
5.6.2
Local Outlier Factor (LOF)
5.7
Tree Based Approaches
5.7.1
Isolation Forest
5.8
Dimension Reduction Based Approaches
5.8.1
Principal Component Analyses (PCA)
5.9
References
References
Published with bookdown
Time Series with R
References