the modern age, the written sources are rapidly increasing. A growing number of this data is related to the texts containing the feelings and opinions of the users. Thus reviewing and analyzing the emotional texts have received a particular attention in the recent years. In this paper, a system that is based on a combination of the cognitive features and the deep neural network, gated recurrent unit, is proposed. Five basic emotions used in this approach are anger, happiness, sadness, surprise, and fear. A total of 23,000 Persian documents by an average length of 24 are labeled for this research work. The emotional constructions, emotional keywords, and emotional POS are the basic cognitive features used in this approach. On the other hand, after pre- processing the texts, the words of the normalized text are embedded by the Word2Vec technique. Then a deep learning approach is followed based on this embedded data. Finally, the classification algorithms such as Naïve Bayes, decision tree, and support vector machines are used in order to classify the emotions based on the concatenation of the defined cognitive features and the deep learning features. 10-fold cross-validation is used in order to evaluate the performance of the proposed system. The experimental results show that the proposed system has achieved an accuracy of 97%. The result of the proposed system shows the improvement of several percent in comparison with the other results achieved by GRU and cognitive features in isolation. At the end, studying other statistical features and improvin
