Replication Package

----- Overview -----Thank you for coming this appendix.This page provides a replication package to you for our each research question.<br>- (RQ1) To what degree are reviews linked?- (RQ2) Why are reviews being linked?- (RQ3) To what degree can link categories be automatically recovered?- (RQ4) To what degree do linked reviews impact code review analytics?<br>----- Description -----<br>Each research question has scripts and data files. RQ1 includes the scripts that compute our quantitative analysis and extract review linkage graphs in our subject systems. RQ2 includes the data that shows our qualitative analysis results (i.e., manually analyzed links). RQ3 has a script that performs multi-class classification for five link categories. Finally, RQ4 has scripts that empirically analyze the impact of reviewer recommenders and outcome predictors.<br>[Environment] MongoDB, Python3.4 R 3.5.1<br>[RQ1]python3.4 linkTrendAnalysis_msr2019.py<br>Rscript plotLinkTrend_msr2019.Rpython3.4 statistics_numReviews_numReviewers.py<br>[RQ2]mr_fse2019.csv<br><br>[RQ3]python3.4 multi_Class_Classification_msr2019.py 1000<br>[RQ4]extract_comments_reviewers.ipynbpredictMultiCategories.ipynboverlappedReviewerAnalysis.ipynbcHRev_recommendation.ipynbrecommender_performance.ipynb<br>outcome_features_extraction.ipynbidenticalOutcomeRate.ipynboutcome_prediction.ipynb<br><br>