No abstract

The object of the research is modern online services and machine learning libraries for predicting the probability of the bank client's consent to the provision of the proposed services. One of the most problematic areas is the high unpredictability of the result in the field of banking marketing using the most common technique of introducing new services for clients – the so-called cold calling. Therefore, the question of assessing the probability and predicting the behavior of a potential client when promoting new banking services and services using cold calling is particularly relevant. In the course of the study, libraries of machine learning methods and data analysis of the Python programming language were used. A program was developed to build a model for predicting the behavior of bank customers using data processing methods using gradient boosting, regularization of gradient boosting, random forest algorithm and recurrent neural networks. Analogous models were built using cloud machine learning services Azure ML, BigML and the Auto-sklearn library. Data analysis and prediction models built using Python language libraries have a fairly high quality – an average of 94.5 %. Using the Azure ML cloud service, a predictive model with an accuracy of 88.6 % was built. The BigML machine learning service made it possible to build a model with an accuracy of 88.8 %. Machine learning methods from the Auto-sklearn library made it possible to obtain a model with a higher quality – 94.9 %. This is due to the fact that the proposed libraries of the Python programming language allow better customization of data processing methods and machine learning to obtain more accurate models than free cloud services that do not provide such capabilities. Thanks to this, it is possible to obtain a predictive model of the behavior of bank customers with a fairly high degree of accuracy. It is worth noting that in order to make a prediction (forecast), it is necessary to study the context of the task, process the data, build various machine learning algorithms, evaluate the quality of the models and choose the best of them.

To ensure the security of information technology, cryptographic information protection tools are used, in particular block and stream encryption algorithms with a symmetric key. Reliability and cryptographic strength of cryptoalgorithms is provided by the properties of the applied primitives. For example, non-linear substitutions (S-boxes) are used as the main component of modern symmetric ciphers. Therefore, generation of substitutions is an important scientific task directly related to the security of information technology and improvement of modern symmetric ciphers. The paper investigates the properties of iterative algorithms for generating non-linear substitutions and special cost functions, which play a decisive role in the heuristic search for S-boxes with the required properties. We consider the cost function of the WCF (Cost Function of the content of the Walsh-Hadamard spectrum) and optimize its parameters. The obtained optimization results in combination with the Hill Climbing iterative search algorithm can reduce significantly the number of iterations. In particular, we show that for a substitution search with a non-linearity of 104, on average, we reduce the computational complexity of generation by more than 20%. In addition, it is possible to increase the success rate of the heuristic search. In particular, for the selected settings, in 100% of cases, a beaktive S-box with a non-linearity of 104 was found.

No abstract

The object of research is the components of an intelligent system for searching information in electronic repositories of unstructured documents, which based on the ontologies of the subject area. One of the most problematic areas is the processing and analysis of information contained in electronic repositories of unstructured documents. There are considered the some possibilities of increasing the efficiency of information processing. In the course of the study, using the method in which ontologies comprise sets of terms presented in it. In addition, the ontological set also includes information about subject areas, areas of definitions, etc. There are obtained the sequence of defining the conceptual representation of an intelligent search system based on ontological components. There are presented the composition of the ontological system model. There are described the main functional components of the system for intelligent processing of information about electronic documents. The proposed approaches for identifying the component components of the ontological model of the search system have a lot of features. This is due to the fact that the search system model must have a set of properties: integrity, coherence, organization, integrability, mobility. Ontologies which representing the basic concepts of the domain in a format available for automated processing in the form of a hierarchy of classes and relationships between them allow automated processing. The using of ontologies in the role of an intermediary between the user and the search process, between the search process and the search system that can facilitate the solution of a number of complex and non-standard tasks of information retrieval (for example, the automation of the search process). It is possible to solve the problem of knowledge representation for displaying information relevant to user requests, as well as to solve the problems of filtering and classifying information. Compared to similar well-known search systems, this provides such advantages as creating a common terminology for software agents and users, protecting the information store from total overflow and errors, as well as solving the issue of information aging.