In this course, you learn about data exploration, feature creation, and feature selection for time sequences. The topics discussed include binning, smoothing, transformations, and data set operations for time series, spectral analysis, singular spectrum analysis, distance measures, and motif analysis.

Learn how to

• Create time series data from transactional or time-stamped data.
• Decompose time series into components of systematic variation.
• Perform spectral analysis to identify cycles in the data.
• Calculate distance measures between series to mitigate collinearity and rank-order features.
• Perform a motif analysis to identify repeating sub-sequences in series.

Who should attend

Analysts with a quantitative background as well as domain experts who would like to augment their time-series tool box

Formats available	Duration
Classroom:	2 days
Live Web Classroom:	4 half-day session(s)	System Requirements
e-Learning:	14 hours/180 day license	System Requirements

Before taking this course, you should be comfortable with basic statistical concepts. You can gain this experience by completing the Statistics 1: Introduction to ANOVA, Regression, and Logistic Regression course. Familiarity with matrices and principal component analysis are also helpful but not required.

This course addresses SAS/ETS, SAS Studio, SAS Visual Forecasting software.

Calculation of Time Series Characteristics

• Accumulation, exploration, and binning.
• Identifying signal and decomposing time series.
• Feature creation and DATA step-like functionality in the TSMODEL procedure.

Distance Measures

• Understand basic ideas in sequence distance measures.
• Apply derived distance measures on candidate input variables to quantify collinearity and correlation with the target.
• Explain symbolic representations of sequences and the usefulness of SAX encoding.

Spectral Analysis and Singular Spectrum Analysis

• Define and discuss analyses in frequency domain.
• Use estimates of the spectral density to identify cycles in the data.
• Explain singular spectrum analysis.
• Apply singular spectrum analysis to identify components of systematic variation in the data such as trends and cycles.

Motif Analysis

• Define and discuss motif analysis and its applications.
• Perform motif analysis on a series, and use the discovered primary motif to score a new sequence.