Local weighted Averaged 2-Dependence Estimator
Proceedings of the 10th Hellenic Conference on Artificial Intelligence, 2018
Despite the extended research that has been made all these years over the classification task exp... more Despite the extended research that has been made all these years over the classification task exploiting algorithms that are based on Bayes theory, the combination of recent semi-naive Bayesian approaches with the well-known concept of Local learning has not highly been scrutinized. We propose a Local variant of Averaged 2-Dependence Estimator (A2DE) as a reliable and high accurate classifier. Our claims are supported by large scale experiments that contain 42 public datasets and 9 different algorithms, including some state-of-the-art, such as the Naive Bayes classifier, and the standalone A1DE and A2DE algorithms for exporting useful conclusions about the proposed learner.
Uploads
Papers by Slobodan Šegrt