Closed
Description
openedon Jun 11, 2020
Feature Description
In time series data, seasonality is the presence of variations that occur at specific regular intervals less than a year, such as weekly, monthly, or quarterly. With the support of seasonality and seasonality decomposition, we can improve a list of operations on time-series data:
- Anomaly Detection
- Forcasting
- and more
We propose to provide
- Seasonality Detection Support for Time-Series Data based on fourier analysis. PR 5231
- Seasonality Decomposition for Time-Series Data based on STL.
a. First, we support decomposition with Anomaly Detection PR 5202
b. Second, separate seasonality decomposition as a individual API as a Transformer
Detail API Proposal
- DetectSeasonality
/// <summary>
/// Obtain the period by adopting techniques of spectral analysis. which is founded by
/// the fourier analysis. returns -1 means there's no significant period. otherwise, a period
/// is returned.
/// </summary>
/// <param name="catalog">The detect seasonality catalog.</param>
/// <param name="input">Input DataView.The data is an instance of <see cref="Microsoft.ML.IDataView"/>.</param>
/// <param name="inputColumnName">Name of column to process. The column data must be <see cref="System.Double"/>.</param>
/// <param name="seasonalityWindowSize">An upper bound on the largest relevant seasonality in the input time-series.
/// When set to -1, use the whole input to fit model, when set to a positive integer, use this number as batch size.
/// Default value is -1.</param>
/// <returns>The detected period if seasonality period exists, otherwise return -1.</returns>
public static int DetectSeasonality(this AnomalyDetectionCatalog catalog, IDataView input, string inputColumnName, int seasonalityWindowSize = -1)
- Seasonality Decompose
Add two optional parameters to existing DetectEntireAnomalyBySrCnn API:
-- period: Seasonality Period (either from user or auto-detected by the DetectSeasonality API.
-- deseasonalityMode: Median, Average, STL.
public static IDataView DetectEntireAnomalyBySrCnn(
this AnomalyDetectionCatalog catalog,
IDataView input, string outputColumnName,
string inputColumnName,
double threshold = 0.3,
int batchSize = 1024,
double sensitivity = 99,
SrCnnDetectMode detectMode = SrCnnDetectMode.AnomalyOnly,
int period = 0,
SrCnnDeseasonalityMode deseasonalityMode = SrCnnDeseasonalityMode.Stl)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
