oviktad

Oviktad is a Swedish adjective meaning unweighted. It is used to describe data, calculations, or models in which all components are treated without weights. The term is the opposite of viktad (weighted) and appears in statistics, mathematics, and computer science within Swedish-language literature. It denotes equal treatment of elements rather than differential emphasis according to assigned weights.

In statistics and data analysis, unweighted procedures apply equal importance to all observations. For example, the unweighted mean computes the simple average by summing values and dividing by the number of observations, whereas the weighted mean assigns different importance via weights. In sampling and survey analysis, unweighted estimates do not adjust for varying probabilities of selection or nonresponse, while weighted estimates use weights to better reflect the target population.

In graph theory and network analysis, an unweighted graph assigns the same significance to all edges, so distances are measured by the number of edges on a path rather than by edge weights. In regression and machine learning, unweighted methods refer to formulations that do not apply observation-specific weights; ordinary least squares is a common unweighted approach, with weighted least squares representing the weighted counterpart.

Limitations: using unweighted methods can introduce bias if elements differ in importance or representation. Weighting is often employed to correct for sampling design, unequal probabilities, or post-stratification, and to reflect varying significance of observations, features, or connections.