US 10,891,218 B2
Automatic pre-detection of potential coding issues and recommendation for resolution actions
George H. Champlin-Scharff, Reading, MA (US); Timothy B. Snow, Westford, MA (US); and Sabrina Yee, Norfolk, MA (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Jul. 1, 2019, as Appl. No. 16/459,292.
Application 16/459,292 is a continuation of application No. 15/843,515, filed on Dec. 15, 2017, granted, now 10,372,592.
Application 15/843,515 is a continuation of application No. 15/367,088, filed on Dec. 1, 2016, granted, now 9,928,160, issued on Mar. 27, 2018.
Application 15/367,088 is a continuation of application No. 14/028,048, filed on Sep. 16, 2013, granted, now 9,519,477, issued on Dec. 13, 2016.
Prior Publication US 2019/0324886 A1, Oct. 24, 2019
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 9/44 (2018.01); G06F 11/36 (2006.01); G06F 8/70 (2018.01); G06F 9/451 (2018.01); G06N 20/00 (2019.01); G06F 8/30 (2018.01); G06F 11/00 (2006.01); G06N 5/04 (2006.01)
CPC G06F 11/368 (2013.01) [G06F 8/30 (2013.01); G06F 8/70 (2013.01); G06F 9/451 (2018.02); G06F 9/454 (2018.02); G06F 11/008 (2013.01); G06F 11/3664 (2013.01); G06F 11/3684 (2013.01); G06N 5/04 (2013.01); G06N 20/00 (2019.01)] 12 Claims
OG exemplary drawing
 
1. A method for automatic pre-detection of potential software product impact and recommendation for resolutions, the method comprising the steps of:
extracting, by a computer processor, symbols from an input phrase expressed in natural language containing a description of an untested software project change;
searching, by a computer processor, using Natural Language Processing with one or more machine learning models, natural language notations in a repository of historical software project changes;
selecting, by a computer processor, in advance of testing of the untested software project change, one or more test cases to employ to regression test and validate changes to the untested software project change according to the Natural Language Processing search and one or more effects attributed to matching the historical software project changes, the input phrase extracted symbols, and the natural language notations and wherein the input phrase comprises a proposed change to the untested software project change;
predicting, by a computer processor, a software error to be induced by programming changes corresponding to the proposed change;
suggesting, by a computer processor, the one or more selected test cases to a user via at least one computer output selected from the group consisting of a user interface device, a user-readable report, and a digital record stored in computer memory, and wherein the suggesting includes the predicted software error; and
responsive to receipt of user approval of the suggested one or more test cases, updating, by a computer processor, the one or more machine learning models.