Network Intrusion Detection (IDS) mechanism is a primary requirement in the current fast growing network systems. Data Mining and Machine Learning (DM-ML) approaches are widely used for network anomaly detection during the past few years. Machine learning based intrusive activity detector is getting popular. However, they produce a high volume of false alarms. One of the main reasons for generating false signals is redundancy in the datasets. To resolve this problem, an efficient feature selection is necessary to improve the intrusion detection system performance. For this purpose, here we use Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Artificial Bee Colony (ABC), K-Nearest Neighbors (KNN) and Support Vector Machine (SVM). The three abovementioned algorithms are used to select the most relevant feature set for identifying network attacks, KNN and SVM algorithms are used as classifiers to evaluate the performance of these feature selection algorithms. The standard NSL-KDD dataset is used for training and testing in this study. We used different metrics to determine which of these algorithms provide a better overall performance when they are used for feature selection in intrusion detection. Our experiments show that PSO, ACO and ABC algorithms perform better than other approaches in feature selection. Feature selection based on ABC provides 98.9% of accuracy rate and 0.78% false alarm with KNN algorithm as the classifier, which is the best result among the examined algorithms.