költségérzékeny

költségérzékeny is a Hungarian term that literally translates to "cost-sensitive." It is most commonly used within the fields of machine learning and data mining, where the goal is to build predictive models that take into account the varying consequences or costs associated with different types of classification errors. In this context, a model is trained not merely to maximize accuracy but to minimize the overall expected cost, which can be defined by a cost matrix or by a more complicated utility function. For example, in medical diagnosis, a false negative (missing a disease) often carries a higher cost than a false positive (an unnecessary alarm), and a költségérzékeny approach would weight these outcomes accordingly.

Historically, cost-sensitive learning emerged in the 1990s as researchers began to recognize that many real-world problems involve asymmetric costs. In addition to classification, the concept has been extended to regression, clustering and ranking problems. Practical implementations include cost-sensitive decision trees, weighted support vector machines, and ensemble methods that explicitly incorporate cost information during both training and decision making. In many industrial and scientific applications—such as fraud detection, quality control, and resource allocation—the use of a költségérzékeny framework has led to more economically efficient outcomes.

Because the Hungarian language often adopts international technical terms with the addition of diacritical marks, költségérzékeny has become a standard term in Hungarian literature on data science. It is typically used interchangeably with its English equivalent, cost-sensitive, and serves as a key descriptor for models that are tailored to minimize domain-specific penalties rather than generic error rates.