The Effect of Ensemble Learning Models on Turkish Text ClassificationDue to rapid development of the Internet and related technologies, the amount of text-based content generated through Internet applications is increasing from day to day. Since text-based content is unstructured, accessing and managing this data is almost impossible. Consequently, there is a need for automatic text classification process. Text mining is a discipline in the Data Mining field and offers algorithms in order to perform text classification. The main objective of text classification is forming a learning model by using a training data set with pre-defined categories and placing data with unknown categories into correct categories. Different text classification algorithms such as decision trees, Bayesian classifiers, rule-based classifiers, neural networks, k-nearest neighbor classifier, support vector machines and ensemble learning methods exist in the literature. In this study, the effect of ensemble learning models on Turkish text classification was evaluated. A publicly available data set named TTC-3600 which consists of 3600 news collected from 6 news portals was selected. Text classification process was performed on TTC-3600 data set by using 4 base classification algorithms Naïve Bayes, Support Vector Machine, K-Nearest Neighbor, J48 Decision tree and their Boosting, Bagging and Rotation Forest ensemble learning models. The experimental results shows that ensemble learning models generally give more accurate results by increasing the success of base classifiers.