classificatieaccuracy

Classification accuracy is a metric that measures the proportion of correct predictions made by a classifier on a data set. It is computed as the number of correctly labeled instances divided by the total number of instances. In multiclass problems, the overall accuracy is the fraction of samples for which the predicted label matches the true label. This simple measure is widely used to provide a quick assessment of a model’s performance.

Computation and related concepts: Accuracy is typically estimated on a held-out test set or via cross-validation. For a given split, it equals the mean of the indicator that a prediction is correct across all samples. The confusion matrix provides a detailed view, from which per-class accuracy, macro-averaged accuracy, and micro-averaged accuracy can be derived. Accuracy can be misleading on imbalanced data, where it may be driven by the majority class. In addition, accuracy does not reveal the confidence of predictions or the calibration of predicted probabilities, so it is often complemented by metrics such as precision, recall, F1-score, ROC AUC, or calibration measures.

Usage and interpretation: Accuracy is intuitive and convenient for quick comparisons, but its interpretation depends on the data set characteristics and the task. It is useful to report alongside other metrics and the confusion matrix to identify common error patterns. When presenting results, it is common to include the evaluation procedure (train-test split or cross-validation), as well as any confidence intervals to convey uncertainty.