Please enable JavaScript to view this site.

Process Designer

Navigation: Modules > JobMind > Usage of the JobMind powered by Parashift systemactivity > Function group - Data extraction per document class

Master Data Recognition

Scroll Prev Top Next More

 

The logic for recognizing supplier and recipient numbers is implemented in JobRouter®.
Parashift only provides the raw data for the algorithm and has no influence on the actual recognition of this information.

Requirements

Master data is required for the recognition of supplier and recipient numbers, which must first be provided to the system.
Please refer to the chapter on Reference data.
If you want to access externally stored reconciliation data, please refer to the subchapter External reference data.

How it works

The detection algorithm uses the extracted data from Parashift. You will find the relevant fields in the following table.

Parameter

Fields

Vendor number

Name

ID

Tax ID (Sender)

pp-tax-id-sender

IBAN

pp-banking-details-iban

Company name

pp-sender-address-company-name

Street

pp-sender-address-street

House number

pp-sender-address-house-number

Postal code

pp-sender-address-postal-code

City

pp-sender-address-city

Recipient number

Name

ID

Tax ID (Receiver)

pp-tax-id-receiver

Company name

pp-receiver-address-company-name

Street

pp-receiver-address-street

House number

pp-receiver-address-house-number

Postal code

pp-receiver-address-postal-code

City

pp-receiver-address-city

The fields listed in the “Fields” column are used to identify the supplier or recipient number.
If one or more of these fields are missing, they are not taken into account in the algorithm.

Please note: If errors from this list are missing in your selected document class, the recognition may not achieve the desired accuracy.

Algorithm and results

The algorithm always returns the candidate with the highest confidence in the output parameters.

To understand how the algorithm recognizes the vendor or recipient numbers, you will find a detailed list of the candidates and the system's confidence in recognizing the data record in the JobActivity log file. The log is only generated if the JobActivity service is running in “Info” mode.

Extracted Values from OCR:

 Name: Office Depot GmbH

 Address: –Soldnerstraße 12 68219 Mannheim

Vendor Matches:

Match #1 (Returned Vendor):

Id: 888888 | Total Confidence: 68.97%

 Address with a confidence of '68.97%': Besselstraße 26 68219 Mannheim

 

2024-11-28 16:31:47.920 +01:00 [INF] Recipient Search Request

Extracted Values from OCR:

 Name: JobRouter Industries AG

 Address: Besselstraße 26 68219 Mannheim

Recipient Matches:

Match #1 (Returned Recipient):

Id: 101 | Total Confidence: 76.67%

 Name with a confidence of '52.38%': JobRouter AG

 Address with a confidence of '92.86%': Besselstrasse 26 68219 Mannheim

Match #2:

Id: 4017 | Total Confidence: 59.37%

 Name with a confidence of '42.86%': JobRouter Democloud

 Address with a confidence of '70.37%': Jobstrasse 26 68219 Mannheim

Match #3:

Id: 3010 | Total Confidence: 34.38%

 Address with a confidence of '34.38%': Bergfeldstrasse 9 83607 Holzkirchen

There is one entry per document in the log with the division between vendor recognition (Vendor) and recipient recognition (Recipient).

Log-Entry

Explanation

Extracted Values from OCR

All data received from Parashift is listed here. The data is used from the fields as described above in “Implementation”.

Vendor Matches & Recipient Matches

Up to 10 candidates are listed in descending order of confidence.

Candidates below 30% will not be considered.

Id

Id of the data record from the master data database

Total Confidence

Weighted sum of the confidence from the address components

... with a confidence of

Levenshtein distance between value from Parashift and master data information.