Field parameters
During training, each word in a labeled field is attributed to a particular class in the neural network. When searching for fields on a new document (for example, one not used in training), the neural network will attribute each word in the new document to one or more classes, with varying degrees of confidence. Additional neural network features will then be used to combine the classified words into fields. These algorithms can be managed using the parameters listed below.| Parameter | Description |
|---|---|
Void FieldName( String ) | The name of the field trained in the Deep Learning activity in the “Group name.Field name” format. |
Void NetName( String ) | The name of the Deep Learning activity. |
Void BreakOnSeparators( Logic ) | If set to False, words separated by vertical separators will be allowed in hypotheses. If set to True, words separated by vertical separators will be excluded from hypotheses. The default value is True. |
Void MinWordCount( Int ) | The minimum number of words allowed in a hypothesis. The default value is 1. |
Void MaxWordCount( Int ) | The maximum number of words allowed in a hypothesis. The default value is 1. |
Void UseBestWordClass( Logic ) | If set to True, each word in a document will be attributed to the neural network class that corresponds to the best-fitting trained field. If set to False, each word in a document will be attributed to all classes with confidence values greater than or equal to MinWordClassConfidence. The default value is False. |
Void StartFromBestWordClass( Logic ) | If set to True, hypotheses will be generated based only on the neural network class that corresponds to the best-fitting field for the first word. If set to False, hypotheses will be generated based on all the classes of the first word. The default value is True. |
Void MinWordClassConfidence( Real ) | The minimum confidence level required for a word to be considered part of a field. The default value is 0.7. |
Void MaxStartWordClassConfidence( Real ) | The maximum confidence level required for a word to be considered part of a field. The default value is 1. The confidence level of a word for belonging to a field must be lower than this value. |
Void ConfidentClassThreshold( Real ) | The threshold for a reliable confidence level. The default value is 0.9. |
Void MinLinkConfidence( Real ) | The minimum confidence level required to combine words into a hypothesis. The default value is 0.5. |
Void MinAvgLinkConfidence( Real ) | The minimum average confidence level required to make a word part of a hypothesis. The default value is 0.5. |
Void MinLinkConfidenceForUncertainWords( Real ) | The confidence level required to make a word part of a hypothesis even if its MinWordClassConfidence is not met. The word’s ConfidentClassThreshold value must not equal that of any other class. The default value is 0.95. |
Table parameters
ProductItemsRepeatableGroup
| Parameter | Description |
|---|---|
Void NetName( String ) | The name of the Deep Learning activity. |
Void MinLineClassConfidence( Real ) | The minimum line confidence required for a line to be considered a table row. The default value is 0.5. |
Void MinColumnCountInRow( Int ) | The minimum number of columns that must be found within a table row. The default value is 2. |
Void CurrencyNames( StringArray ) | Names of currencies that may appear next to numbers. Specify this parameter to improve detection of columns containing amounts. This parameter will be deprecated in future versions. |
Void UnitsOfMeasure( StringArray ) | Units of measurement that may appear next to amounts of goods. Set this parameter to improve detection of columns containing amounts. This parameter will be deprecated in future versions. |
ProductItemsColumn
| Parameter | Description |
|---|---|
Void FieldName( String ) | The name of the field that is trained by the Deep Learning activity. |
Void MinClassConfidence( Real ) | The minimum confidence level required for a word to be considered as belonging to the column. It can be a decimal number from 0 to 1. The default value is 0.2. |
Void IsDecimal( Logic ) | Indicates whether the column consists of decimal numbers. The default value is False. |
Void IsDescription( Logic ) | Indicates whether the column contains descriptions and is not suitable for dividing the table into rows. The default value is False. |
Void IsMultiline( Logic ) | Indicates whether the column contains multiline cells. The default value is False. |
Void IsUnfixed( Logic ) | Indicates whether the column’s position within the table is unfixed. The default value is False. |
Void DivideRowsByColumn( Logic ) | Indicates whether the column is used to divide the table into rows. The default value is True. |