MS SQL SERVER: Time series algorithm
1 like2,549 views
The document provides an overview of Microsoft Time Series algorithms for forecasting continuous values like sales over time. It discusses creating time series models in SQL Server that use a key time column to identify time steps in the data. Examples are given for creating a DMX query to define a mining structure and model using the Microsoft Time Series algorithm, specifying parameters and predictable attributes. The query is executed to generate the model and structure on the server.
![DMXCREATE MINING MODEL statement:CREATE MINING MODEL [<Mining Structure Name>] ( <key columns>, <predictable attribute columns> ) USING <algorithm name>([parameter list]) WITH DRILLTHROUGHThe code defines the key column for the mining model, which in the case of a time series model uniquely identifies a time step in the source data. The time step is identified with the KEY TIME keywords after the column name and data types. You can have multiple predictable attributes in a single mining model. When there are multiple predictable attributes, the Microsoft Time Series algorithm generates a separate analysis for each series:](https://image.slidesharecdn.com/microsofttimeseriesalgorithm-100711090622-phpapp02/85/MS-SQL-SERVER-Time-series-algorithm-8-320.jpg)
![DMX(Altering the Query )In Query Editor, copy the generic example of the CREATE MINING MODEL statement into the blank query.Replace the following: 1. [mining model name] with [Forecasting_MIXED] 2. <key columns> with [Reporting Date] DATE KEY TIME, [Model Region] TEXT KEY The TIME KEY keyword indicates that the ReportingDate column contains the time step values used to order the values. The TEXT and KEY keywords indicate that the ModelRegion column contains an additional series key.](https://image.slidesharecdn.com/microsofttimeseriesalgorithm-100711090622-phpapp02/85/MS-SQL-SERVER-Time-series-algorithm-10-320.jpg)
![DMX(Altering the Query ) Replace the following:3. < predictable attribute columns> ) with: [Quantity] LONG CONTINUOUS PREDICT, [Amount] DOUBLE CONTINUOUS PREDICT )4. USING <algorithm name>([parameter list]) WITH DRILLTHROUGH with:USING Microsoft_Time_Series(AUTO_DETECT_PERIODICITY = 0.8, FORECAST_METHOD = 'MIXED') WITH DRILLTHROUGH](https://image.slidesharecdn.com/microsofttimeseriesalgorithm-100711090622-phpapp02/85/MS-SQL-SERVER-Time-series-algorithm-11-320.jpg)
![DMX(Altering the Query )The complete statement should now be as follows:CREATE MINING MODEL [Forecasting_MIXED] ( [Reporting Date] DATE KEY TIME, [Model Region] TEXT KEY, [Quantity] LONG CONTINUOUS PREDICT, [Amount] DOUBLE CONTINUOUS PREDICT ) USING Microsoft_Time_Series (AUTO_DETECT_PERIODICITY = 0.8, FORECAST_METHOD = 'MIXED') WITH DRILLTHROUGH On the File menu, click Save DMXQuery1.dmx As.In the Save As dialog box, browse to the appropriate folder, and name the file Forecasting_MIXED.dmx.](https://image.slidesharecdn.com/microsofttimeseriesalgorithm-100711090622-phpapp02/85/MS-SQL-SERVER-Time-series-algorithm-12-320.jpg)
MS SQL SERVER: Time series algorithm
- 1. Microsoft Time Series Algorithm
- 2. overviewUnderstanding Microsoft Time Series algorithmTime Series ScenariosDMXCreating the QueryAltering the Query Executing the Query
- 3. Microsoft Time Series algorithmThe Microsoft Time Series algorithm provides regression algorithms that are optimized for the forecasting of continuous values, such as product sales, over time.A time series consists of a series of data collected over successive increments of time or other sequence indicator. A time series model can predict trends based only on the original dataset that is used to create the model.
- 4. Microsoft Time Series algorithmshows a typical model for forecasting sales of a product in four different sales regions over time.Historical information appears to the left of the vertical line and represents the data that the algorithm uses to create the model.Predicted information appears to the right of the vertical line and represents the forecast that the model makes
- 5. Time Series ScenariosPerforming a Simple ForecastThis step involvescreating a model, showing it some data, and asking it for predictions of future values.Predicting Interdependent SeriesThe Microsoft implementation finds relationships where they exist between series and will use these relationships in forecasting.Marking a series as INPUT indicates that it cannot be forecasted and will be considered only as to its impact on other series. Marking a series as PREDICT_ONLY, on the other hand, indicates that the series can be forecasted, but will not be considered by other series.
- 6. Time Series ScenariosUnderstanding Your Time SeriesYou can explore the individual patterns that are being used for prediction and see if the previous sales have a larger impact on future behavior.you can get more descriptive rules about your data, because the Time Series algorithm is based on the decision tree implementation used in the Microsoft Decision Trees algorithm.
- 7. DMXThe main element that differentiates a mining structure used for time series from other structures is the inclusion of a KEY TIME column. The KEY TIME content type indicates that a column is a KEY as well as the time slice representing the row.
- 8. DMXCREATE MINING MODEL statement:CREATE MINING MODEL [<Mining Structure Name>] ( <key columns>, <predictable attribute columns> ) USING <algorithm name>([parameter list]) WITH DRILLTHROUGHThe code defines the key column for the mining model, which in the case of a time series model uniquely identifies a time step in the source data. The time step is identified with the KEY TIME keywords after the column name and data types. You can have multiple predictable attributes in a single mining model. When there are multiple predictable attributes, the Microsoft Time Series algorithm generates a separate analysis for each series:
- 9. DMX(Creating the Query )To create a new DMX query in SQL Server Management StudioOpen SQL Server Management Studio.In the Connect to Server dialog box, for Server type, select Analysis Services. In Server name, type LocalHost, or the name of the instance of Analysis Services that you want to connect to for this lesson. Click Connect.In Object Explorer, right-click the instance of Analysis Services, point to New Query, and then click DMX.Query Editor opens and contains a new, blank query.
- 10. DMX(Altering the Query )In Query Editor, copy the generic example of the CREATE MINING MODEL statement into the blank query.Replace the following: 1. [mining model name] with [Forecasting_MIXED] 2. <key columns> with [Reporting Date] DATE KEY TIME, [Model Region] TEXT KEY The TIME KEY keyword indicates that the ReportingDate column contains the time step values used to order the values. The TEXT and KEY keywords indicate that the ModelRegion column contains an additional series key.
- 11. DMX(Altering the Query ) Replace the following:3. < predictable attribute columns> ) with: [Quantity] LONG CONTINUOUS PREDICT, [Amount] DOUBLE CONTINUOUS PREDICT )4. USING <algorithm name>([parameter list]) WITH DRILLTHROUGH with:USING Microsoft_Time_Series(AUTO_DETECT_PERIODICITY = 0.8, FORECAST_METHOD = 'MIXED') WITH DRILLTHROUGH
- 12. DMX(Altering the Query )The complete statement should now be as follows:CREATE MINING MODEL [Forecasting_MIXED] ( [Reporting Date] DATE KEY TIME, [Model Region] TEXT KEY, [Quantity] LONG CONTINUOUS PREDICT, [Amount] DOUBLE CONTINUOUS PREDICT ) USING Microsoft_Time_Series (AUTO_DETECT_PERIODICITY = 0.8, FORECAST_METHOD = 'MIXED') WITH DRILLTHROUGH On the File menu, click Save DMXQuery1.dmx As.In the Save As dialog box, browse to the appropriate folder, and name the file Forecasting_MIXED.dmx.
- 13. After a query is created and saved, it needs to be executed to create the mining model and its mining structure on the server. In Query Editor, on the toolbar, click ExecuteA new structure named Forecasting_MIXED_Structure now exists on the server, together with the related mining model Forecasting_MIXED.DMX(Executing the Query )
- 14. Visit more self help tutorialsPick a tutorial of your choice and browse through it at your own pace.The tutorials section is free, self-guiding and will not involve any additional support.Visit us at www.dataminingtools.net