Data Mining Practice Test

Data that does not have known outcome values for the attribute you wish to predict are described as unlabeled data. This kind of data is what you work with in unsupervised learning, where there’s no target variable to guide learning, so you look for patterns, clusters, or structure rather than predict a specific label. The other terms fit different roles: training data includes the known outcomes used to teach a model; labeling is the process of adding those outcomes; testing data is used to evaluate a model and also comes with known outcomes for comparison. Scoring means assigning a numerical score or ranking to instances, which is an operation done after a model is built and doesn’t by itself describe the absence of outcome values.