CSKH02.2019 - (Abstract) - Recently, the number of incidents related to Web applications, due to the increase in the number of users of mobile devices, the development of the Internet of things, the expansion of many services and, as a consequence, the expansion of possible computer attacks. Malicious programs can be used to collect information about users, personal data and gaining access to Web resources or blocking them. The purpose of the study is to enhance the detection accuracy of computer attacks on Web applications. In the work, a model for presenting requests to Web resources, based on a vector space model and attributes of requests via the HTTP protocol is proposed. Previously carried out research allowed us to obtain an estimate of the detection accuracy as well as 96% for Web applications for the dataset KDD 99, vectorbased query representation and a classifier based on model decision trees.