Network attacks face problems of weakened security on computer networks, hindering the normal flow of services due to unauthorized access to data. Such attacks comprise phishing, malware and viruses, password guessing, and forced entry on passwords and the network devices. Such attacks can result in loss of business revenue as the company’s services may take time to be restored, with the costs incurred in recovering from such an attack and future legal consequences for the company being huge. In order to address these problems, the active learning method based on the K-means clustering technique of machine learning (ML) is proposed for the identification of malicious attacks. ALM with K-means clustering is proposed to be used for this proposed malicious detection technique for keeping a check on both internal and external attacks. Rule-based classification is one of the processes by which data is classified in accordance with the number of rules. The proposed method attains a commendable accuracy of 98.5%, which is more than the existing approaches.