CSKH-02.2020. Abstract—Currently, the number of information leaks caused by internal violators has increased. One of the possible channels for information leaks is the transmission of data in encrypted or compressed form, since modern DLP (data leakage prevention) systems are not able to detect signatures and other information related to confidential information in such data. The article presents an algorithm for classifying sequences formed by encryption and compression algorithms. An array of frequencies of occurrence of binary subsequences of length N bits was used as a feature space. File headers or any other contextual information were not used to construct the feature space. The presented algorithm has shown the accuracy of classification of the sequences specified in the work 0.98 and can be implemented in DLP systems to prevent the transmission of information in encrypted or compressed form.