Browse Prior Art Database

Enabling Automatic Extraction And Traceability Of Global Requirements From Current Engagement Documents

IP.com Disclosure Number: IPCOM000241610D
Publication Date: 2015-May-17
Document File: 4 page(s) / 108K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for enabling automatic extraction and traceability of global requirements from current engagement documents using machine learning models.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 4

Enabling Automatic Extraction And Traceability Of Global Requirements From Current Engagement Documents

A typical IT engagement starts with a short-lived and low-cost business value analysis.

An accurate estimate and contract for an implementation phase of a project is created based on use cases, scenarios, and high-level requirements of a customer. Business value analysis inputs include current engagement documents (CED), namely meeting notes, emails, requirements/specification documents, and other information provided by customers based on customer requirement. A business value assessment output is a contract that typically includes a list of global requirements used for risk assessment, an associated estimation, and pricing.

Manual methods are used to analyze artifacts to extract global requirements from current engagement documents, by consulting engagement corpus and requirements from past engagements in a Report Mining Server (RMS).

Disclosed is a method and system for enabling automatic extraction and traceability of global requirements from current engagement documents using machine learning models that are trained using previous engagement corpus and global requirements as training data. Each previous engagement results in two kinds of artifacts - past engagement corpus and requirements. The past engagement corpus are natural text documents such as, email, notes, etc. The past engagement corpus may be maintained in a Content Management System (CMS) or may be managed by the RMS. Requirements are well defined structured artifacts managed in a RMS.

In accordance with the method and system, a requirement unit contains sections in the original document that represent or describe an engagement requirement. The engagement requirement is created based on the engagement corpus. Requirement unit boundaries are determined by text structures such as headings, paragraphs, and bullets and are linked / traced to an engagement document with a traceability link maintained by the RMS. Requirements units that need to be classified are contained in the RMS and possess traceability links to the engagement requirement.

The method and system uses traceability links that are traced from requirement units to engagement requirements to global requirements managed in the RMS, to create ground truth and training data for its statistical models. Engagement requirements and global requirements are identified and linked to each other manually. Traceability links are used to create training sets for a natural language classifier, where requirements units are mapped to global requirement.

The method and system identifies global requirements and potentially traces global requirements to plans, tasks, estimates and change sets in an Application Lifecycle Management (ALM) system. In addition, risk assessment and estimation is provided for known global r...