Implement confusion_matrix(data) for binary classification with labels 0 and 1.
data is a list of [y_true, y_pred] pairs (actual label, predicted label), one per observation.
Return a matrix as a list of lists of integer counts, laid out as
y_true = 1, y_pred = 1y_true = 1, y_pred = 0y_true = 0, y_pred = 1y_true = 0, y_pred = 0Note this row order (actual positives first) differs from scikit-learn's [[TN, FP], [FN, TP]].