Electronic template generation, data extraction and response detection

摘要

A file receiver receives an electronic structure file that includes structure-file data associated with a spatial arrangement and detects a content object for processing that includes content-object data. A file transformation engine transforms the structure-file data from the structure file into an electronic record. A rendering engine renders an image of the transformed structure-file data arranged in the spatial arrangement. An interface engine detects an input corresponding to specification of a position of a data segment. A parsing engine defines a segment-position specification indicative of the position. A template engine generates an electronic template that associates an identifier of the data segment with the segment-position specification and associates the electronic template with a template identifier. A record classifier determines that the content object corresponds to the template identifier. The parsing engine further extracts, using the segment-position specification, a portion of the content-object data that corresponds to the data segment.

主权项

1. A content management system for extracting portions of content objects for semi-automated answer identifications, the content management system comprising:
a file receiver that detects a set of assessment content objects for processing, each assessment content object of the set of assessment content objects including content-object data arranged according to a spatial arrangement, each assessment content object of the set of assessment content objects being associated with an identifier corresponding to a student; a record classifier that, for each assessment content object of the set of assessment content objects, determines that the assessment content object corresponds to an identifier of a template for a particular assessment, the template identifying a segment-position specification that at least partly defines an area in which an answer for a question is to have been provided; a parsing engine that, for each assessment content object of the set of assessment content objects:
identifies a portion of the content-object data as corresponding to an answer area for the question based on the segment-position specification and the spatial arrangement; andextracts the portion of the content-object data; a response classifier that:
evaluates the portion of the content-object data to identify an answer indicated in the portion of the content-object data;determines an evaluation quality metric reflecting a confidence in the identification of the answer; anddetermines whether a quality criterion is satisfied based on the evaluation quality metric; and an interface engine that:
facilitates a presentation that includes:
multiple portions of content-object data, the multiple portions of content-object data having extracted from multiple assessment content objects from the set of assessment content objects, each portion of the multiple portion being associated with a determination that the quality criterion is not satisfied; anda tool that receives external input identifying an answer indicated in one or more portions of the multiple portions; andreceives an input corresponding to an identification of an answer indicated in a portion of the one or more portions; wherein the response classifier further stores, upon receiving the input, the answer in association with the identifier of the student associated with the assessment content object from which the portion was extracted.