System and method for identifying optimal test cases for software development
AssigneeAtos Syntel, Inc.
InventorsPadmanabhan Balasubramanian, Indumathi Balaji, Yogeswari Radha
Life of this patent
Estimating the term…
Abstract
A test case identification system that may tokenize natural language input and correlate the resultant set of tokens to a plurality of test cases for identification of a test case from the plurality of test cases. The identified test cases may be scored based upon historical information, quantification metrics and number of search token matches in order to identify a test case to a user. The results of the tokenization, correlation and identification may be stored in a master library for utilization in future identification efforts.
Claims20
A computerized system for identifying a test case stored in a test case repository, the system comprising:
- a computer comprising a processor and a memory;
- a natural language processing module stored in the memory of the computer and, when executed by the processor, configured to:
- receive an input containing a requirement;
- apply a natural language processing rule to the input to generate a first output;
- and
- tokenize the first output into a set of tokens comprising a first search token;
- an analytics module stored in the memory of the computer and, when executed by the processor, configured to:
- cluster the first search token with a first quantification metric, whereby the first quantification metric is retrieved from a master database based upon a correlation between the first search token and a first persistent term associated with the first quantification metric;
- cluster the first search token with a first test case, whereby the first test case is retrieved from the test case repository based upon a correlation between the first search token and a tokenized description associated with the first test case;
- calculate a score of the first search token based upon the first quantification metric;
- calculate a score of the first test case based upon the score of the first search token;
- and
- a reporting module stored in the memory of the computer and, when executed by the processor configured to generate a report comprising the score of first test case.
A method of identifying a test case from a test case repository, the method comprising:
- receiving an input containing a requirement;
- applying a natural language processing rule to the input to generate a first output;
- tokenizing the first output into a set of tokens comprising a first search token;
- clustering the first search token with a first quantification metric, whereby the first quantification metric is retrieved from a master database based upon a correlation between the first search token and a first persistent term associated with the first quantification metric;
- clustering the first search token with a first test case, whereby the first test case is retrieved from the test case repository based upon a correlation between the first search token and a tokenized description associated with the first test case;
- calculating a score of the first search token based upon the first quantification metric;
- calculate a score of the first test case based upon the score the first search token;
- and
- generating a report comprising the score of first test case.
Dependencies are inferred from the claim text.