US 12,124,800 B2 · Computerized systems and methods for hierarchical structure parsing and building

Abstract

Disclosed are systems and methods for a computerized framework that provides a document structure parsing system for requirement engineering documents, where the logical structure of the text is not available, and is to be rebuilt based on the raw textual content. The framework approaches the build of the logical structure according to two phases. The first phase involves creating a list of list of text snippets from the raw text, where sequence labeling is adopted to re-segment and merge initially segmented text snippets. The second phase involves the framework executing computerized techniques including embedding adaptation approach, a hierarchy structure rebuilt algorithm, and a requirement text selection strategy to rebuild the hierarchy structure.

Claims20

Dependent claims

A computer-implemented method for document structure rebuilding, comprising:

identifying, by a processor, an electronic requirement document (ERD) comprising text;
analyzing, by the processor, the ERD to determine a listing of text snippets within the text;
performing adaptive embedding, by the processor, on the listing of text snippets, the adaptive embedding comprising executing at least one neural model to determine an embedding for each text snippet;
executing, by the processor, a document recursive structure building algorithm based on the determined embeddings to determine a tree structure for the ERD;
determining, by the processor, a set of similarity values of each embedding, the set of similarity values corresponding to a prototype factor and a requirement factor, the prototype factor representing a group of text snippets, the requirement factor representing a sequence of text snippets;
clustering, by the processor, each embedding based set of similarity values;
and
rebuilding, by the processor, the ERD based on the clustered embeddings and determined tree structure.

5 more elements

A non-transitory computer-readable storage medium tangibly encoded with computer-executable instructions, that when executed by a processor, perform a method for document structure rebuilding comprising steps of:

identifying an electronic requirement document (ERD) comprising text;
determining based on an analysis of the ERD, a listing of text snippets within the text;
determining an out-of-vocabulary (OOV) ratio that indicates a value of OOV words in a respective text snippet;
executing at least one neural model on the listing of text snippets to determine an embedding for each text snippet;
executing a document recursive structure building algorithm based on the determined embeddings for the listing of text snippets, the execution causing a determination of a tree structure for the ERD;
determining a prototype similarity value and a requirement similarity value for the determined embeddings;
determining a maximal margin relevance (MMR) value based on the prototype similarity value and the requirement similarity value;
and
rebuilding the ERD based on the MMR determination and determined tree structure.

6 more elements

Dependencies are inferred from the claim text.

Computerized systems and methods for hierarchical structure parsing and building

Life of this patent

Computerized systems and methods for hierarchical structure parsing and building

Life of this patent

Abstract

Claims20