In table analysis, each candidate is evaluated using the criteria listed below.
| Table property | Range of Results |
|---|---|
| Fraction of default columns identified | 0.0 – 1.0 |
| Fraction of table columns mapped | 0.0 – 1.0 |
| Fraction of entries found in cells which require entries | 0.0 – 1.0 |
| Fraction of neighboring cells without overlap (column view) | 0.0 – 1.0 |
| Fraction of neighboring cells without overlap (row view) | 0.0 – 1.0 |
| Fraction of correct formats | 0.0 – 1.0 |
For each table candidate, the total significance is determined as the average of the results per property. All table candidates, which reach a certain threshold for the total significance, are transferred to the workdoc, ordered by the total significance. By default, the candidate with the highest significance is used for data extraction.
Note: During table analysis, field-level validation settings defined through the user interface
are ignored. However, you can modify the evaluation settings using scripts.