Presentation is loading. Please wait.

Presentation is loading. Please wait.

ML Integrated Software Testing First American India Private Limited

Similar presentations


Presentation on theme: "ML Integrated Software Testing First American India Private Limited"— Presentation transcript:

1 ML Integrated Software Testing First American India Private Limited
Akhil Gopal V, QA Engineer Uday Kiran Kuramana, Associate Manager Tanuja Rajpurohith, QA Engineer First American India Private Limited

2 Machine Learning concepts and Algorithms into Software Testing.
ABSTRACT With the adoption of Agile process in software development, time to market has decreased drastically. In order to ensure the quality of deliverables, we need to upgrade our Test Strategy iteratively. This presentation gives an insight as to how Test strategy can be strengthened by incorporating Machine Learning concepts and Algorithms into Software Testing.

3 PROBLEM STATEMENT Software testing is getting limited to Predefined and known test data Selection of Test Cases and Data is only depending upon Tester’s perspective Not aware about the probability of occurrence/pattern of any Real Time Data Difficulty in understanding how customer is using the application Difficulty in dealing with Big Data No selection criteria for Regression Test Cases being executed iteratively

4 APPROACH

5 Approach Explained Initially, metadata of frequent combinations/traces is extracted from real time database using suitable Association algorithm, like the Apriori Mapping correlations to an excel sheet which serves as input for automation system Selecting/Designing Test Suites as per the relevant test data identified and automating the test execution Building an automation system with a suitable algorithm which facilitates learning of the system on how to try out combinations of Test Data and Test Cases iteratively with respect to the real time customer behavior

6 Concept Explained The process of applying a machine learning method in Software Testing can be broken into 5 main steps. Data Analysis – Descriptive statistics and carry out exploratory data analysis. Data Preparation – Clean, reformat, and aggregate data. Algorithm Testing and Evaluation – Compare algorithms and evaluate the results of each. Selecting the appropriate algorithm Model Fine-Tuning – Make small tweaks to model, in order to improve its performance. Implementation – Get the model into production and accuracy within time

7 BENEFITS Greater Agility with high quality
Better Customer Segmentation and Accurate Lifetime Value Predictions Simplifies Time-Intensive Testing methodologies Helps in selection of Regression suites for successive iterations based on real time application usage The approach leverages ROI of the software being delivered Optimization of manual/automation resources and allocation to right process

8 References & Appendix › Conferences

9 Author Biography Uday Kiran Kuramana works as a Associate Manager at First American India Private Limited, Bangalore and has 10 years of experience in Software Industry Akhil Gopal Venugopal works as a QA Engineer at First American India Private Limited, Bangalore and has 2 years of experience in Software Industry Tanuja Rajpurohith works as a QA Engineer at First American India Private Limited, Bangalore and has 2 years of experience in Software Industry

10 Thank You!!!


Download ppt "ML Integrated Software Testing First American India Private Limited"

Similar presentations


Ads by Google