InTDS ArchivebyTam D Tran-The·May 12, 2023Data-Oriented Programming with PythonA recap on Data-Oriented Programming by Yehonathan Sharvit illustrated with Python examples (instead of JavaScript and Java)A response icon10A response icon10
InTDS ArchivebyTam D Tran-The·Aug 25, 2022How to Evaluate Any Tabular Synthetic DatasetA series of mechanisms and tests one can use to evaluate any tabular synthetic dataset, from resemblance and utility perspectivesA response icon2A response icon2
InTDS ArchivebyTam D Tran-The·Jul 7, 2022How cross-validation can go wrong and how to fix it: feature selection use case with sample codeThere are multiple nuances when designing a cross-validation procedure that might make the obtained results less robust or even wrongA response icon1A response icon1
InTDS ArchivebyTam D Tran-The·May 20, 2022The wrong and right way to approximate Area Under Precision-Recall Curve (AUPRC)There are many ways to summarize the AUPRC, but not all are of equal meritA response icon1A response icon1
InTDS ArchivebyTam D Tran-The·Mar 2, 2022Performance Curve: More Intuitive than ROC/PRC and Less Assumptive than Threshold MetricsAn evaluation method for binary classifiers that combines the best of both worldsA response icon2A response icon2
InCodeXbyTam D Tran-The·Mar 2, 2022Feature Selection with a Random VariableA trick to make your decision of choosing top n most important features less arbitrary
InTDS ArchivebyTam D Tran-The·Feb 3, 2022Sample Size Planning for Interrupted Time Series Design in Health CareITS is one of the strongest quasi-experimental designs. Properly planning for the study is arguably more important than the analysis…
InTDS ArchivebyTam D Tran-The·Nov 29, 2021Precision-Recall Curve is more informative than ROC in imbalanced dataWhen the negative class is more prevalent and there is low value in true-negative predictions, Precision-Recall curve is preferred over…A response icon2A response icon2
InTDS ArchivebyTam D Tran-The·Nov 24, 2021Credibility of Data ScienceFor data science to be effectively applied in the real world, its problem of credibility should not be taken lightly.