Advanced Reports
Advanced Reports are informative tools that assist administrators in analyzing and optimizing the parameters used by Transact in the Document Assembly and Extraction modules.
Transact generates reports for batches that have the Review or Validation module present in the workflow. The following diagram illustrates a sample workflow that contains both the Review and Validation modules.
Reporting monitors Transact performance over time and captures the information about the correction process. This tool provides insight into adjustments that the administrator can make to the system's classification, extraction rules, and thresholds. The system can be fine-tuned for maximum performance and increased ROI.
The following Advanced Reports are available:
-
Document Correction Report
-
Classification Accuracy Report
-
Classification Correction Details Report
-
Separation Accuracy Report
-
Separation Correction Details Report
-
Unnecessary Review Report
-
False Positive Report
-
Extraction Correction Report
-
Field Correction Report
-
Field Correction Details Report
Advanced Reports can be exported as a PDF or Excel file.
Data archival scripts
Overview
This topic describes how to manually acquire scripts for Archiving Data from Advanced Reports. This can be done to improve performance of ETL scripts when a large amount of data is accumulated in database tables.
Required configuration
There are two scripts provided below, based on whether the customer database is SQL Server or MariaDB. In both cases, the user must replace the following placeholders with values from their database:
-
@@ARCHIVE_REPORT_DB_NAME@@: Name of the Archive Database that will be created when script is run.
-
@@REPORT_DB_NAME@@: Name of the current Reports database that is used by Transact.
These scripts need root/administrator access because a new database must be created.
Document Correction Report
Document Correction Reports are top-level reports that provide the user with all document types and the number of manual corrections that the batch instance operator made during the Document Assembly phase.
The Document Correction Report allows the user to identify where a specific problem with the document type may have occurred. Examples of possible document type issues include:
-
The operator modified the document type during the Classification process (Classification Correction)
-
The operator split or merged different documents (Separation Correction)
-
Operator input was not needed to modify documents in Classification (Unnecessary Review)
-
The operator modified the document type during the Validation process (False Positive)
The following components are included in the reports display panel for the Document Correction Report:
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
The Document Correction Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit Type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
Id of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class. Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Document Correction Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class Example: Mailroom Automation |
|
Document Type |
Types of documents defined in batch class |
|
Classification Correction Count |
Number of classification corrections made by the reviewer. Classification correction is the change in document type in the Review module. |
|
Separation Correction Count |
Number of page separations made by the reviewer. Separation correction is the change in page count due to Split/Merge/Delete. |
|
Unnecessary Review Count |
Number of unnecessary reviews made by the reviewer. It is the number of document types or pages changed by user during review. |
|
False Positive Count |
Change of document type during validation when batch did not stop for review, in the Review module. |
|
Total Correction Count |
The total number of all types of corrections made by the reviewer or validator. |
Report body
The Document Type Vs Correction Count Chart widget is a bar chart that graphically represents the type of correction count made in each document.
Navigation and drilldowns
The Document Correction Report is the default report displayed in Advanced Reports. The user can access the Classification Accuracy Report, Separation Accuracy Report, Unnecessary Review Report, and the False Positive Report from the data grid.
From the Document Correction Report, the user can identify the areas where the issues related to document types are located. Based on this report, it can be noted that unnecessary reviews and false positives are the simplest and most efficient fixes.
Classification Accuracy Report
The Classification Accuracy Report is a mid-level report that aggregates corrections made to document types within Classification with respect to all batches run. This gives the user a measure of accuracy for each document type of every batch class.
The following components are included in the reports display panel for the Classification Accuracy Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Classification Accuracy Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class. Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Classification Accuracy Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Document Type |
Types of documents defined in batch class. |
|
Accuracy |
The accuracy with which each document type of every batch class is classified. It Is measured in percentage. |
Report body
Document Type vs Accuracy Chart is a bubble chart that graphically represents the percentage of accuracy with which each document type of every batch class in classified. The accuracy percentage is represented in the form a bubble. Size of the bubble indicates the frequency of corrections made by the user.
Navigation and drilldowns
From this report, the user can access the Classification Correction Report from the data grid as well as the displayed chart. The user can also return to the parent report (Document Correction Report) using the link provided on the top left corner of the report.
Classification Correction Details Report
Filters
The Classification Correction Details Report is a low-level report detailing corrections made to document types within Classification. A classification correction is the result of a change in the document type of documents in Review state by the user.
The following components are included in the reports display panel for the Classification Correction Detail Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
-
Filters
Classification Correction Details Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
Id of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Classification Correction Detail Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Batch Instance Id |
Id assigned to each batch instance. |
|
Start Date |
Date when the batch was picked up by Transact for execution. |
|
Document Id |
ID of the document. |
|
Old Document Type |
Document type identified by Transact. |
|
New Document Type |
New document type assigned by reviewer. |
|
Threshold |
The minimum confidence for the Batch to identified under specified document. |
|
Confidence |
The score with which document is identified by Transact. |
|
Page Count |
Number of pages in each document. |
Report body
The user can return to the parent report (Classification Accuracy Report) using the link provided on the top left corner of the report. Utilizing the two lower level reports gives the user an idea on what needs to be changed within Transact configurations.
Separation Accuracy Report
The Separation Accuracy Report is a mid-level report which aggregates corrections made to page counts in Review module with respect to all batches run. This gives the user a measure of accuracy for each document type of every batch class.
The following components are included in the reports display panel for the Separation Accuracy Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Separation Accuracy Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
ID of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Separation Accuracy Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example:BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class Example: Mailroom Automation. |
|
Document Type |
Types of documents defined in batch class. |
|
Separation Accuracy |
The accuracy with which each page of the document type of every batch class is separated, measured in percentage. |
Report body
The Document Type versus Accuracy chart is a bubble chart that graphically represents the percentage of accuracy with which each document type of every batch class is separated. The accuracy percentage is represented in the form a bubble. Size of the bubble indicates the frequency of corrections made by user.
Navigation and drilldowns
The user can access the Separation Correction Report from the data grid and displayed chart or return to the parent report (Document Correction Report) using the link provided on the top-left corner of the report.
Separation Correction Details Report
The Separation Correction Details Report is a low-level report detailing corrections made to page counts within Review module. Separation correction is the result of a split, merge, or deletion of a page in a document during Classification. These result in the modification of the number of pages within a document.
The following components are included in the reports display panel for the Separation Correction Details Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Separation Correction Details Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
ID of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class. Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Separation Correction Details Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class ID |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Batch Instance ID |
ID assigned to each batch instance. |
|
Start Date |
Date when the batch was picked up by Transact for execution. |
|
Document ID |
ID of the document. |
|
Document Type |
The document to which the batch belongs. |
|
Old Page Count |
The number of pages identified by Transact. |
|
New Page Count |
The number of pages identified and updated by the reviewer. |
Navigation and drilldowns
The user can return to the parent report (Separation Accuracy Report) using the link provided on the top left corner of the report. Using the two lower-level reports gives the user an idea on what needs to be changed within Transact configurations.
False Positive Report
False Positive is the opposite of an Unnecessary Review. It occurs when a document goes through Classification without any issue, such as the document confidence being higher than the threshold for the document type, but the document is changed by an operator in Validation.
This issue is caused by thresholds that are set too low.
By looking at the False Positive Report, the administrator can determine if the threshold should be increased for a specific document type. If the document type for the document was changed even when the confidence was higher than the threshold, it was a false positive.
The following components are included in the reports display panel for the False Positive Report:
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
False Positive Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
Id of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the False Positive Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Batch Instance ID |
Id assigned to each batch instance. |
|
Start Date |
Date when the batch was picked up by Transact for execution. |
|
Document Id |
Id of the document. |
|
Old Doc Type |
Document type identified by Transact. |
|
New Doc Type |
New document type selected by validator. |
|
Confidence |
The score with which document is identified by Transact. |
|
Old Doc Type Threshold |
Document threshold value of old document type. |
Report body
Document Type vs Confidence Chart is a floating bar chart that graphically represents the confidence and threshold value on Y axis and document type on the X axis; smaller the bar better is the threshold assigned.
Navigation and drilldowns
The user can return to the parent report (Document Correction Report) using the link provided on the top left corner of the report.
Unnecessary Review Report
A document is labeled as an unnecessary review if it goes into review when the document confidence is below threshold, and the operator simply confirms the auto-classified document type.
This issue is caused by thresholds that are set too high. By looking at the Unnecessary Review Report, the user can determine a more appropriate threshold to be set in Transact.
If batches being reported in the Unnecessary Review Report have a confidence level of 20.00 and a threshold closer to 50.00, the threshold for the document type in Transact should be decreased for more accurate processing.
The following components are included in the reports display panel for the Unnecessary Review Report:
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Unnecessary Review Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
Id of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The Unnecessary Review Report Data displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Batch Instance Id |
ID assigned to each batch instance. |
|
Start Date |
Date when the batch was picked up by Transact for execution. |
|
Document Id |
ID of the document. |
|
Document Type |
Document type identified by Transact. |
|
Confidence |
The score with which document is identified by Transact. |
|
Threshold |
The minimum confidence for the batch to identified under specified document. |
Report body
Document Type versus Confidence Chart is a floating bar chart that graphically represents the confidence and threshold value on Y axis and document type on the X axis.
Navigation and drilldowns
The user can return to the parent report (Document Correction Report) using the link provided on the top left corner of the report.
Extraction Correction Report
Extraction Correction Reports are top level reports that provide the user with all document types and the number of manual corrections during Extraction. Users are able to easily identify which document types (if any) require further investigation.
The following components are included in the reports display panel for the Extraction Correction Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Extraction Correction Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
Id of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class. Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of Documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Extraction Correction Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Document Type |
Document type to which the corrected field belongs to. |
|
Total Field Count |
Total number of fields for which values are extracted by Transact. |
|
Average Confidence |
Average of extracted OCR confidence values for all fields. This data is pulled from the batch.xml |
|
Field Change Count |
Sum of field changes for all batches of this field, this document type and this batch class. |
|
Change Ratio |
Change ratio=(field change count/number of fields). |
Report body
Document Type versus Change Ratio chart is a bubble chart that graphically represents the change ratio (that is, the number of fields changed to total fields extracted) in the document type. The size of the bubble indicates the frequency of corrections made by the user.
Navigation and drilldowns
From this report the user can access the Field Correction Report from the data grid as well as the displayed chart. From the Extraction Correction Report, the user can easily identify where the issues corresponding to document types are located.
Field Correction Report
Field Correction Reports are mid-level reports that provide the user with all fields of all document types and the number of manual corrections during Extraction. Users are able to easily identify which fields of document types (if any) require further investigation.
The following components are included in the reports display panel for the Extraction Correction Report:
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Extraction Correction Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
ID of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of documents defined in each Batch Class. Default Value: All |
Data grid
The data grid of the Field Correction Report displays information about executed batches in tabular format.
|
Column Name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Document Type |
Document type to which the corrected field belongs to. |
|
Field Name |
Field name for which the value is corrected |
|
Total Count |
Sum of documents with this document type in all batches of this field, this document type and this batch class |
|
Average Confidence |
Average of extracted OCR confidence values for all fields. This data is pulled from the batch.xml |
|
Field Change Count |
Sum of field changes for all batches of this field, this document type and this batch class |
|
Change Ratio |
Change ratio = (field change count/number of fields) |
Charts
Field Name versus Change Ratio chart graphically represents the change ratio of the field value versus the field. The frequency of corrections done by the user is represented in the form a bubble. Size of the bubble indicates the frequency of corrections made by the user.
Navigation and drilldowns
From this report, the user can access the Field Correction Details Report from the data grid and the displayed chart. The user can also return to the parent report (Extraction Correction Report) using the link provided on the top left corner of the report.
From the Extraction Correction Report, the user can easily identify where the issues corresponding to the field types are located.
Field Correction Details Report
Field Correction Details Reports are low-level reports that provide the user with all fields of all document types and the number of manual corrections during Extraction. Users are able to easily identify which fields of document types (if any) require further investigation.
The Field Correction Details Report provides field-by-field information on what field data was extracted from a document (the original value) and what an operator entered during Validation (the corrected value).
There is a Batch Class Name column in the Field Correction Details table in Advanced Reports.
The rp_data_corrections table stores information about document corrections at manual steps in workflow. This table stores information that could track changes in the document name, page count, confidence score within the two review stages in the workflow. This table stores pre- and post-state of document correction data in form of JSON object its column.
The following components are included in the reports display panel for the Field Correction Details Report.
-
Filters
-
Data grid
-
Report body
-
Drilldowns and navigation
Filters
Field Correction Details Report filters allow the user to specify the parameters to generate reports. Depending on the value selected in filters, the result is maximized or minimized. View the available filters in the following table.
|
Filter name |
Options |
Submit type |
Description |
|---|---|---|---|
|
Start Date |
Calendar |
Submit button |
Start date of the period for which executed batch report is to be generated. |
|
End Date |
Calendar |
Submit button |
End date of the period for which executed batch report is to be generated. |
|
Batch Class |
All Batch Classes present in Transact |
Submit button |
ID of the Batch Class. Example: BC1 for Mailroom Automation Template Batch Class Default Value: All |
|
Document Type |
All Types of Documents defined in each Batch Class |
Submit button |
Types of documents defined in each Batch Class. Default Value: All |
|
Field Name |
All field names that can be extracted |
Submit button |
Name of the fields that can extracted. Default Value: All |
Data grid
The data grid of the Field Correction Details Report displays information about executed batches in tabular format.
|
Column name |
Description |
|---|---|
|
Batch Class Id |
ID of the batch class. Example: BC1 for Mailroom Automation Template Batch Class |
|
Batch Class Name |
Name of the batch class. Example: Mailroom Automation |
|
Batch Instance Id |
Id assigned to the batch instance. |
|
Batch Start Date |
Date when the batch is picked up by Transact for execution. |
|
Document Id |
Id of the document. |
|
Document Type |
Document type whose field is corrected. |
|
Field Name |
Field name for which the value is corrected. |
|
Field Old Value |
Value of the field extracted by Transact. |
|
Field New Value |
Value updated by the validator for the field. |
| Extraction C3454672ewqstyuixcvbnm,1onfidence |
The confidence with which value is extracted. |
Navigation and drilldowns
The user can return to the parent report (Field Correction Report) using the link provided on the top left corner of the report. From the Extraction Correction Report, the user can easily identify where the issues corresponding to the field types are located.