Document coding provides a systematic, standardized organization of documents for efficient data retrieval. The objective of Coding is to identify and capture relevant content from the collection of images. It creates
database (s) of choice, based on the fields pertinent to specific case requirements. Data Fields typically include document date, title, type, Author, Recipient, Copyee as well as Characteristics such as Marginalia, Foreign Language, Handwritten etc.
Logical document boundaries and Attachment Range relationships are critical to the validity and usefulness of a database. In logical unitization we have to use key variables to distinguish documents from each other and also define attachment relationships among contiguous documents. This aspect of Bibliographic coding can also be customized to allow for Email threads to be coded individually or bundle similar documents based on title, date etc. as sometimes required for issue based coding.
Our coders will need to be trained to meet the unique needs of each project ranging from Standard Bibliographic Coding to highly complex multi-tiered in-text and subjective data capture. Names-in-text, marginalia names and keywords sometimes need to be captured to facilitate in-depth coding for easy retrieval of all pertinent records.
Logical document boundaries and Attachment Range relationships are critical to the validity and usefulness of a database. In logical unitization we have to use key variables to distinguish documents from each other and also define attachment relationships among contiguous documents. This aspect of Bibliographic coding can also be customized to allow for Email threads to be coded individually or bundle similar documents based on title, date etc. as sometimes required for issue based coding.
Our coders will need to be trained to meet the unique needs of each project ranging from Standard Bibliographic Coding to highly complex multi-tiered in-text and subjective data capture. Names-in-text, marginalia names and keywords sometimes need to be captured to facilitate in-depth coding for easy retrieval of all pertinent records.
Coding is the most subjective phase of document processing. We will need to establish multifaceted quality control measures to ensure both the accuracy and consistency of all coded data, to meet the industry standard of at least 97% overall database accuracy.
A specifically designed data capture software, is used to enter data following the specific set of guidelines created for each project. Coding supervisors monitor workflow and ensure qualitative and quantitative goals are being met. As unusual situations develop, the project manager consults the client's designated contact representative. When new decisions are made regarding coding standards and specifications, the staff receives regular training updates.
The team leaders will review a sample of each coder's work on a daily basis. A specialized monitoring system will allow them to select random samples of coded document records. The team leaders will compare each field from the selected records against the original documents. If they find inconsistencies, they will return the documents to the coders. Coders will be required to correct their own errors and check their previous work for any likely errors.
Team leaders will perform 100% quality control during the initial coding period, comparing every field from every record coded against the original document. This level of quality control will continue until a coder consistently meets accuracy requirements. Quality checks will continue throughout the project as part of daily quality monitoring.
Using a custom designed program, the quality controller will review each new entry in the personal name and organization table, correcting misspellings, ensuring formatting conformance with project specifications, merging multiple and inconsistent entries of the same organization, etc.
The quality assurance team will check coded data for inconsistencies called "pattern errors". For example, the team might search all patent documents within a batch for an entry in the Patent Number field. If no patent number is present, the quality assurance specialist, using a split screen (records on one side and images on the other) will simultaneously look at the image of the document record and enter the missing patent number (or in some cases may change the document type). For projects requiring a Title field entry, quality assurance team will sort the Titles of all documents in a batch. On those without an entry, the team will display the images and enter Titles.
After extracting data, the project manager will review it one final time to ensure correctness of formatting and output.
© 2007 DelhiNet LPO. All rights reserved