Coding system for high data volume
AssigneeInternational Business Machines Corporation
InventorsEugene Wallach, Aviad Zlotnick
Life of this patent
Estimating the term…
Abstract
A method for automated coding of a text phrase relative to a catalog of codes. The method includes finding a plurality of the codes that are candidates for coding of the phrase and identifying a category to which one or more of the candidate codes belong. The phrase is conveyed together with the one or more candidate codes in the identified category to a human operator specialized in the identified category, for verification by the operator of one of the candidate codes in the category for assignment to the phrase.
Claims26
A method for automated coding of a text phrase relative to a catalog of codes, comprising:
- finding a plurality of the codes that are candidates for coding of the phrase;
- identifying a category to which one or more of the candidate codes belong;
- selecting a human operator specialized in the identified category;
- and
- conveying the phrase together with the one or more candidate codes in the identified category to the selected human operator, for verification by the operator of one of the candidate codes in the category for assignment to the phrase.
A method for automated coding of text phrases made up of words, the method comprising:
- providing a collection of reference phrases and codes respectively assigned to the reference phrases by one or more human operators;
- generating for the words in the reference phrases respective lists of the codes, the list for each word comprising the codes assigned to the reference phrases containing the word;
- receiving an input phrase to which a code is to be assigned;
- processing the input phrase using the lists of codes so as to determine whether one or more candidate codes, selected from the lists, meet a criterion for assignment to the phrase;
- and
- if one of the selected candidate codes meets the criterion, making a record indicative of the assignment of the one of the selected candidate codes to the phrase.
Apparatus for automated coding of a text phrase relative to a catalog of codes, comprising:
- a plurality of coding workstations, adapted to be operated by human operators having respective fields of specialization;
- and
- a coding server, coupled to communicate with the workstations, and operative to find a plurality of the codes that are candidates for coding of the phrase, to identify a category to which one or more of the candidate codes belong, and to convey the phrase together with the one or more candidate codes in the identified category to one of the workstations that is operated by one of the human operators whose field of specialization includes the identified category, for verification by the operator of one of the candidate codes in the category for assignment to the phrase.
Apparatus for automated coding of text phrases made up of words, the apparatus comprising a coding server, adapted to receive a collection of reference phrases and codes respectively assigned to the reference phrases by one or more human operators, to generate for the words in the reference phrases respective lists of the codes, the list for each word comprising the codes assigned to the reference phrases containing the word, to receive an input phrase to which a code is to be assigned, and to process the input phrase using the lists of codes so as to determine whether one or more candidate codes, selected from the lists, meet a criterion for assignment to the phrase.
A computer software product for automated coding of a text phrase relative to a catalog of codes, comprising a computer-readable medium in which program instructions are stored, which instructions, when read by a computer, cause the computer to find a plurality of the codes that are candidates for coding of the phrase, to identify a category to which one or more of the candidate codes belong, to select a human operator specialized in the identified category, and to convey the phrase together with the one or more candidate codes in the identified category to the selected human operator, for verification by the operator of one of the candidate codes in the category for assignment to the phrase.
A computer software product for automated coding of text phrases made up of words, the product comprising a computer-readable medium in which program instructions are stored, which instructions, when read by a computer, cause the computer, upon receiving a collection of reference phrases and codes respectively assigned to the reference phrases by one or more human operators, to generate for the words in the reference phrases respective lists of the codes, the list for each word comprising the codes assigned to the reference phrases containing the word, and upon further receiving an input phrase to which no code has yet been assigned, to process the input phrase using the lists of codes so as to determine whether one or more candidate codes, selected from the lists, meet a criterion for assignment to the phrase.
Dependencies are inferred from the claim text.