Classification systems are methods of grouping and organizing data so that they can be compared with other data. The type of classification system used depends on what the data is intended to measure. Some datasets may use multiple classification systems. Feature selection is the most important step in classification systems and plays an important role in many fields such as pattern recognition, machine learning, signal processing and other data mining. The goal of feature selection is to find the smallest subset of input features and it is widely used in high-dimensional data, such as text processing and classification, which improves classification performance. Therefore, the high dimensions of the text is the most important problem in text classification. This paper presents a new method, based on the cuckoo optimization method, which finds optimal or semi-optimal solutions in polynomial time complexity, and the proposed method is easily implemented using a simple classifier. In order to demonstrate the power of the proposed method, the performance is compared with a genetic algorithm, cuckoo-based optimization method, chi-square, information gain (IG) and another combination of cuckoo and genetic algorithm on the benchmark dataset. The simulation results of this dataset show the superiority of the proposed feature selection.
