Day or week profiles 2.0

Summary

This operator calculates a distribution profile for the underlying observations, e.g. 2.5% of all observations occur between 07:00 and 08:00, 3.5% between 08:00 and 09:00, etc.

Configuration

Input settings of existing table

Name	Value	Opt.	Description	Example
Identifier	System.Object	opt.	Please select those tables by whose content the data is to be grouped. A separate profile will be created for each group.	-
Date + Time (from)	System.DateTime	-	Please select the column that contains the start times (date + time) of your observations.	-
Date + Time (to)	System.DateTime	-	Please select the column that contains the final times (date + time) of your observation.	-
Observations	System.Double	-	This (numerical) column contains the values, which have been observed over a day or a whole week.	-

Settings

Name	Value	Opt.	Description	Example
Apply profiles to	System.Int32	opt.	Enter here a data node to which the calculated profiles are to be applied.	-
Identifier	System.String	opt.	Please select those columns which correspond to the identifiers given above. These columns determine which profile is to be applied to a forecasted observation.	-
Date + Time (from)	System.String	opt.	Please select the column that contains the start times (date + time) of your observations.	-
Observations	System.String	opt.	This (numerical) column contains the forecasted value for a day or a week.	-
Deliver as result	System.String Profile per weekday Profile per weekday/month Profile per weekday/quarter Profile per weekday/quarter Profile per weekday/quarter Profile per week/quarter Profile based on all data Best profile per identifier Best profile per identifier/weekday Best profile across all identifiers Validation results	-	Please select which data should be displayed in the results.	-
Calculation method	System.String Average Median Mode Percentile 99% Percentile 97% Percentile 95% Percentile 90% Percentile 85% Percentile 80% Percentile 75% Percentile 70% Percentile 60%	-	Please select which daily and weekly profiles are to be calculated, e.g. via the mean value, median or percentile of the observations on which they are based.	-
Smoothing	System.String No smoothing Average Linear interpolation	-	After a profile has been calculated a smoothing method can be applied in order to avoid extreme differences between two consecutive estimates. Please select the smoothing method to be applied to the profiles.	-
Validation period	System.String Every month Last month Last 3 months Last 6 months	-	The observations are divided into a validation period and training period, e.g. validation period = the last three months of the data on which the model is based, training period = all other data. Profiles are created on the basis of the training period, and these profiles are used to estimate the data of the training and validation period. The percentage and average errors between the estimated and actual observations are determined for the training and validation period. The smallest validation error will be used to determine the best profile per identifier, per identifier/weekday or the best profile across all identifier.	-
Output the validation result in data nodes	System.Boolean	opt.	If selected, the validation results are displayed in a separate data node.	-
Output error messages/warnings in data nodes	System.Boolean	opt.	If selected, error messages and warnings are displayed in a separate data node.	-