Abstract: An imbalanced classification problem is one in which the distribution of instances across defined classes is uneven or biased in one direction or another. In data mining, the probabilistic ...

Abstract: Automated design of metaheuristic algorithms offers an attractive avenue to reduce human effort and gain enhanced performance beyond human intuition. Current automated methods design ...

General approach for combining modular metaheuristic software frameworks and automatic configuration tools. To discourage the inefficient manual invention and configuration of new metaheuristic ...

In four new studies, researchers found complicated results from experiments on Facebook’s and Instagram’s algorithms, suggesting there was no silver bullet to fixing the platforms. By Mike Isaac and ...