accuracy

Accuracy is the closeness of a measured value or estimate to the true value. It reflects trueness or correctness of a result. Precision, by contrast, is about consistency of repeated measurements; a result can be precise but inaccurate if it misses the true value.

Accuracy is affected by systematic errors (bias) that shift results away from the true value, and by random errors that cause scatter. Improving accuracy involves calibration against reference standards, ensuring traceability, controlling environmental factors, and using appropriate measurement methods. Quantitatively, accuracy can be expressed as absolute error, relative error, or bias, often accompanied by an uncertainty interval.

In statistics and machine learning, accuracy commonly means the proportion of correct predictions in a dataset. This form of accuracy can be misleading with imbalanced classes or different costs of errors; other metrics such as precision, recall, F1 score, or area under the ROC curve are used to provide a fuller picture.

The concept is central in science, engineering, quality control, and data analysis, guiding instrument design, data collection, and evaluation of model performance.