Advanced integrated requirements engineering system for CE-based requirements assessment

• US 5,748,973 A
• Filed: 07/15/1994
• Issued: 05/05/1998
• Est. Priority Date: 07/15/1994
• Status: Expired due to Term

• Current Assignee: George Mason Intellectual Properties, Inc.
• Original Assignee: George Mason University
• Inventors: Evans, Richard P., Palmer, James D.
• Primary Examiner(s): Hayes, Gail O.
• Assistant Examiner(s): Tkacs, Stephen R.
• Application Number:
  US08/275,422
• Time in Patent Office: 1,390 Days
• Field of Search: 364/419.01, 364/419.08, 364/419.13, 364/419.19, 395/12, 395/76, 395/919, 395/922, 395/701, 395/708, 395/751, 395/759, 395/761, 395/792, 395/793
• US Class Current: 704/9
• CPC Class Codes: G06F 40/237 Lexical tools

Abstract

An automated, user-interactive procedure is used to obtain the right statements for computer software from the user in a way in which the statements elicited from the user are free from errors due to ambiguity, conflict, redundancy, inconsistency, and the like without compromise of user intent. The advanced integrated statements engineering system (AIRES) is an integrated environment, supporting and spanning the statements engineering life cycle. Each activity in the statements engineering life cycle is supported by a set of automated CASE tools, which taken together, comprise a suite of CASE tools in the AIRES environment. A primary feature of the assessment framework is a classification ectype (CE), which is a combination of operator controlled rules and tables to represent a category within a classification structure. CE-based statements assessment (CEBRA) provides automated support techniques to "recognize" and thus categorize English text. The AIRES-CEBRA approach applies two CE forms in combination; semantic-based and syntactic-based. The semantic-based CE from includes, as elements of the rules, text attributes such as parts of speech and order and juxtaposition of terms. Successive application of a set of such rules enables the application of a wide variety of recognition techniques including sophisticated linguistic pattern matching rules as well as semantic classification principles. The syntactic CE form enables categorization without constraint of semantic content, considering only term presence or absence. The combination of the two CE forms, semantic and syntactic, significantly improves the effectiveness of the AIRES invention, over and above their significance individually to text categorization.

Claims

1. A user-interactive computer-aided English text categorization method that supports the use of quantified conditions and implemented on a computer having a computer user interface for prompting and receiving user selections and inputs, said method comprising the steps of:

accepting input as English language text, formatting the text for subsequent processing, and assigning identifications (IDs) to sets of the input English language text;

establishing the characteristics of a complete set of text as the basis for the quantifiable conditions, the characteristics including co-occurring terms, pair-wise syntactic similarity coefficients (SCs), combinations of parts of speech;

constructing interactively with prompted operator input classification ectypes (CEs) as text-edited rules and tables to specify the quantified conditions for representation of a category within a classification structure, said CEs being of two forms, word form and integer form CEs, said word form CEs being based on a list of designated words from the input English language text, said integer form CEs being of two types, ID-based and SC-based, ID-based CEs supporting logical operations as a basis for categorizing statements and SC-based CEs using the pair-wise SCs between statements; and

applying constructed CEs to prepare lists of candidates for each category.

2. The user interactive computer aided engineering statements method recited in claim 1 wherein the step of establishing statements characteristics comprises the steps of:

applying a weighted words file to the English language text to de-emphasize or eliminate words in the English language text;

preparation of a master vector of all words or terms remaining in the English language text;

applying weighted files of synonyms, parts-of-speech, and uses relationships to the English language text to apply weighted values for these relationships;

applying operator-selected stemming operations to reduce terms to their basic roots; and

identifying parts of speech and combinations of parts of speech meaning and context, including subject nouns and subject noun phrases, verbs, verb phrases, and subject noun coupled with verb phrases.

3. The user interactive computer aided engineering statements method recited in claim 2 wherein the step of establishing statements characteristics further comprises the step of calculating similarity coefficients (SCs) as a dot product or cosine between each pair of statements.

4. The user interactive computer aided engineering statements method recited in claim 1 wherein the step of constructing CEs includes constructing word form CEs by prompting a user to select among a plurality of files, each of said files establishing quantified conditions to be satisfied including, part of speech, order, synonym and other relationships, tables of conditions such as date of change, degree of juxtaposition of terms, and case sensitivity.

5. The user interactive computer aided engineering statements method recited in claim 1 wherein the step of applying CEs comprises the steps of:

receiving word form CE files and integer form CE files in the form of rules and tables and receiving characteristics files as additional tables generated by said steps of establishing statements characteristics and constructing CEs;

applying the rules and tables of the word form CE files together with additional tables of the characteristics files to generate a word and ID form output in categories defined by applied CEs;

applying the rules and tables of the integer form CE files together with additional tables of the characteristics files and the word and ID form output to generate an integer form output; and

applying, as quantified conditions in variable threshold combinatorials (VTCs) combinations of the members of sub-categories of the text.