Anushree Agrawal, Ph.D

INDIRA GANDHI DELHI TECHNICAL UNIVERSITY FOR WOMEN

− Improved existing changeability prediction model with the design, implementation and validation of novel software ripple effect prediction model. The accuracy of the model evaluated using open-source software application was 92.8%− Learned and applied several supervised ML algorithms such as Neural Network, Bagging, Random Forest and Adaboost for the ripple effect identification problem and compared the performance using various statistical measures and Friedman Test.− Discovered a… Show more − Improved existing changeability prediction model with the design, implementation and validation of novel software ripple effect prediction model. The accuracy of the model evaluated using open-source software application was 92.8%− Learned and applied several supervised ML algorithms such as Neural Network, Bagging, Random Forest and Adaboost for the ripple effect identification problem and compared the performance using various statistical measures and Friedman Test.− Discovered a weight assignment scheme for old and new change commits, to solve the problem of obsolete dependencies and improve prediction accuracy.− Addressed the challenges of class imbalance and irrelevant features for the ML based prediction models.− Implementation of association rule mining in “Ruffle” using Python for change history retrieved from svn and git repositories. Show less

Institute of Engineering and Technology

Activities and Societies: - FITI (Information Technology Society) - Coordinator and volunteer of various college events including literary events Coordinator in ENCORE' 09 and assistant coordinator in ENCORE'08. − Design and implementation of online job search website using Dreamviewer.

Delhi Technological University (Formerly DCE)

Activities and Societies: Software Engineering Society - Proposed software defect prediction model using data derived from source code and logs extracted from software project repositories to identify erroneous classes in software applications.- Developed a tool “CMS” for defect and change data collection from the cvs repository of open-source software projects. − Learned and applied Logistic Regression, Neural Network and Random Forest Machine Learning techniques for software fault prediction and obtained AUC up to 88.2% with ten-fold… Show more - Proposed software defect prediction model using data derived from source code and logs extracted from software project repositories to identify erroneous classes in software applications.- Developed a tool “CMS” for defect and change data collection from the cvs repository of open-source software projects. − Learned and applied Logistic Regression, Neural Network and Random Forest Machine Learning techniques for software fault prediction and obtained AUC up to 88.2% with ten-fold validation.− Came up with the idea of cross-project defect prediction and studied the feasibility using 28 features derived from the source code and Decision Tree Analysis. - Identified relevant features in the dataset using correlation analysis with the AUC of the cross project prediction models up to 90.1%. Show less

Software Developer

Associated with Anushree Agrawal, Ph.D

Laboratory Teaching Assistant

Associated with Anushree Agrawal, Ph.D