SVM-KM: speeding SVMs learning with a priori cluster selection and k-means
Distribution of the number of citations over years.