This project analyzes the impact of Mercedes-Benz Stadium's opening on crime patterns in Atlanta neighborhoods using difference-in-differences methodology.

This analysis evaluates the performance of a binary classification model by implementing core classification metrics from scratch and validating those results against established R packages such as caret and pROC. Using the provided dataset (classification-output-data.csv), I compute and interpret accuracy, error rate, precision, recall/sensitivity, specificity, F1 score, and ROC/AUC to assess model behavior across multiple dimensions. These metrics are essential for understanding how well a model distinguishes between positive and negative classes, especially when considering false positives, false negatives, and threshold-dependent performance. The analysis also includes a manually constructed ROC curve and trapezoidal AUC approximation, followed by comparisons with package-generated outputs to confirm correctness and deepen understanding of classification evaluation methods.

für Gruppe 3B

An Unsupervised Learning Approach Using NHANES Data

assignment