CPC G06F 11/3684 (2013.01) [G06F 11/3664 (2013.01); G06F 11/3688 (2013.01); G06N 20/00 (2019.01)] | 18 Claims |
1. A method implementing software performance testing comprising:
receiving first test sample data having a first volume, the first test sample data representing operation of a software function;
determining a format of the first test sample data to be not numerical;
in response to determining that the format of the first test sample data is not numerical, applying the first test sample data to a first pre-trained machine learning model to produce a plurality of different sized data corpuses, each of the data corpuses being larger than the first volume and representing operation of the software function over a plurality of timescales;
testing the software using the plurality of different sized data corpuses and producing a plurality of results corresponding to the plurality of timescales; and
applying the first test sample data to a second pre-trained machine learning model, different from the first pre-trained machine learning model, and repeating said testing step in response to determining the plurality of results is non-linear.
|