Bulletin of Electrical Engineering and Informatics Vol. 13, No. 3, June 2024, pp. 1897~1912 ISSN: 2302-9285, DOI: 10.11591/eei.v13i3.6003  1897 Journal homepage: http://beei.org Ensemble learning based on relative accuracy approach and diversity teams Mahmoud B. Rokaya 1,3 , Kholod D. Alsufiani 2 1 Department of Information Technology, College of Computers and Information Technology, Taif University, Taif, Saudi Arabia 2 Computer Sciences Program, Turabah University College, Taif University, Taif, Saudi Arabia 3 Department of Computer Science, Faculty of Science, University of Tanta, Tanta, Gharbia, Egypt Article Info ABSTRACT Article history: Received Feb 12, 2023 Revised Oct 14, 2023 Accepted Nov 14, 2023 Ensemble learning, which involves combining the opinions of multiple experts to arrive at a better result, has been used for centuries. In this work, a review of the major voting methods in ensemble learning is explored. This work will focus on a new method for combining the results of individual learners. The method depends on the relative accuracy and diversity of teams. Instead of trying to assign weight to each different trainer, the concept of diversity teams is presented. Each team will vote as one player; however, the individual accuracies of each learner still be implemented. The concept of relaxing parameters that deal with each team as one player is presented. Our experiments demonstrate that traditional ensemble voting methods outperform individual learners. There is a limit to the superiority of the ensemble learner that any ensemble learner cannot go beyond. The proposed voting method gives the same results as the traditional ensemble voting methods, however, a different diversity of the proposed method from the traditional voting method or for different values of the relaxing parameter can be achieved. Keywords: Decision tree Ensemble learning Recurrent neural network Support vector machine Weighted voting This is an open access article under the CC BY-SA license. Corresponding Author: Mahmoud B. Rokaya Department of Information Technology, College of Computers and Information Technology Taif University Taif 21944, Saudi Arabia Email: mahmoudrokaya@tu.edu.sa 1. INTRODUCTION Ensemble learning is defined as “ensemble learning depends on training a set of trainers and then using these trainers for implementing new data through taking a weighted vote of the trainer’s results [1]. There are many combining methods, however, the most known methods are bagging and boosting [2]. Ensemble learning has many approaches and many works looked at considering many aspects. Some works focused on the types of trainers, either homogenous [3], [4] or heterogeneous [5], [6]. Some works looked at the purpose of using ensemble learning either for classifying [7]–[26], clustering [27]–[34], regression [35]–[37], or streaming [38]–[42]. Ensemble learning has many applications in almost all fields. In medicine, some works applied ensemble methods to predict the disease [43]–[50] or to classify the patients in each disease [50]–[56]. Also, ensemble learning is used in medicine for DNA prediction [57] or for DNA imbalanced splice site datasets [58]. Ensemble learning has applications in security [59]–[63]. Social media is not an exception, and many works depend on ensemble learning to achieve better performance or cover multi-data types for one learner [64]–[74]. Ensemble learning is used in commerce [75]–[78] and credit cards [79]–[81]. Image processing is a traditional field for machine learning and ensemble learning, for example [82], [83]. Industry [84]–[94],