Method and system to attribute metadata to preexisting documents
View Patent ↗This disclosure provides a method and system to convert a first digital representation of a document including a first format to a second digital representation of the document including a second format. According to an exemplary method, a preexisting printed assessment is scanned to produce an image file and an educational assessment analysis system captures an image area associated with a question, processes the captured image area to automatically assign metadata associated with an independent assessment creation process to the captured image area, which is subsequently processed to generate an assessment based on the assigned metadata which conforms with a format consistent with the independent assessment creation process.
1. A computer-implemented method of creating an assessment using an educational assessment analysis system including one or more questions to be answered by a student, the method comprising:
a) the educational assessment analysis system acquiring a first digital representation of a preexisting assessment not created with the educational assessment analysis system;
b) the educational assessment system capturing a first area of the first digital representation of the preexisting assessment, the first area selected by a user of the educational assessment analysis system and the first area including text associated with a question to be answered by a student;
c) the educational assessment system assigning metadata to the first area, the metadata including a plurality of one or more predefined user selectable metadata field entries associated with the first area and the plurality of predefined user selectable metadata field entries including metadata field entries generated by a plurality of trained metadata field type classifiers and a plurality of metadata classification category classifiers for each field type classification, each metadata classifier trained using a training set associated with assessments generated by an independent assessment creation process and each metadata field type classifier classifying the first area text as being associated with a single metadata field type and each metadata classification category classifier further classifying the first area text as being associated with two or more classification categories ordered by likelihood and associated with the single metadata field type, and the plurality of predefined user selectable metadata field entries defined by the associated independent document creation process configured to generate assessments based on one of a plurality of predefined assessment models and based on the user inputting the plurality of one or more predefined selectable metadata field entries;
d) the educational assessment system displaying each metadata field populated with a most likely classification category of the ordered classification categories from the metadata classification category classifiers and displaying a user selectable menu listing in order of likelihood other classification categories of the ordered classification categories from the metadata classification category classifiers selectable by a user; and
e) the independent assessment creation process generating a second digital representation of the preexisting assessment based on the first area and the user assigned metadata and processable by the educational assessment analysis system to evaluate questions and associated student answers included in the preexisting assessment,
wherein prior to step e), step a), step b) step c) and step d) are repeated to capture a second area of the first digital representation of the preexisting assessment and assign metadata to the second area, and step d) generates the second digital representation of the preexisting assessment document based on the first area, the first area's assigned metadata, the second area, and the second area's assigned metadata, and,
wherein the captured first area is associated with one of a distinct first layer and first cell, and the metadata is assigned to one of the first layer and first cell and the captured second area is associated with one of a respective second layer and second cell, and the metadata is assigned to one of the second layer and second cell.
2. The computer-implemented method of creating an assessment according to claim 1 , wherein all or part of the metadata is automatically assigned to the first area, step c) comprising:
c1) extracting one or more of numbers, characters, key words and key phrases from the first area, the numbers, characters, key words and key phrases associated with the one or more predefined selectable metadata fields;
c2) accessing the plurality of trained metadata classifiers to determine the plurality of predefined selectable metadata field entries associated with the first area, each of the trained metadata classifiers associated with a respective predefined selectable metadata field entry; and
c3) assigning the determined selectable metadata field entries as metadata associated with the first area.
3. The computer-implemented method of creating an assessment according to claim 2 , wherein the plurality of trained metadata classifiers are derived from one or more heuristics applied to the extracted one or more numbers, characters, key words and key phrases from the first area.
4. The computer-implemented method of creating an assessment according to claim 2 , wherein the plurality of trained metadata classifiers are trained using a machine learning process including a supervised training set.
5. A computer program product comprising a non-transitory recording medium storing instructions, which when executed on a computer causes the computer to perform the method for converting a first digital representation of a first form assessment document to a second digital representation of the form assessment document according to claim 1 .