Search

Predictive psychodiagnostics is the process of forming a predictive assessment of any metric based on the data of psychodiagnostic testing. The formation of a predictive estimate is a classification problem and a variety of algorithmic approaches can be used to solve it – logistic regression, support vector machines, decision trees, neural networks, and many others. The article describes the practice of using algorithms based on decision trees to create models used in solving predictive psychodiagnostics problems, as well as cases in which their application is most relevant. The results obtained using the models built on the basis of the decision tree algorithm using gradient boosting show sufficient predictability and statistical significance in comparison with the methods of traditional psychodiagnostics used to solve personnel assessment problems

The article deals with the architecture and structure of the most effective artificial neural networks (ANNs) in the tasks of network traffic classification: multilayer perceptron and LSTM (long short-term memory) network. Using one of the actual datasets CSE-CIC-IDS2018 trained ANNs and tested the effectiveness of detection of various attacks in the network traffic. An analysis of existing work has been performed. A multiclass classification was performed and the detection effectiveness of separate computer attacks was reported. The effectiveness evaluation focused on three metrics: completeness, correct response rate, and false positive rate. The test classification results were considered both in terms of comparison of numerical characteristics of recognition quality and in terms of class analysis in order to identify which computer attacks (CA) are better identified using multilayer perceptron and which are better identified using LSTM. On the basis of these results, an ANN, which is a hybrid network, is proposed as a promising Intrusion Detection System (IDS).

Currently, unmanned vehicles cannot yet be introduced on public roads, and the person behind the wheel is not predictable and behavior is a multi-parameter task. The article describes a methodology for calculating and evaluating various characteristics of a respondent's psychodiagnostic test using the integrated use of typical driver behavior uncertainty and assessing his condition through machine learning (ML) algorithms. The application of machine learning allows you to reduce the number of test questions and increase the processing speed, since the developed algorithm finds hidden patterns well. A large number of factors affect the driver's condition, and this article is devoted to determining the driver's condition in real time, which reduces the likelihood of an emergency. Based on the developed algorithm, it is possible to conduct an interview and pre-trip control. On the one hand, this makes it possible to improve the quality of the recruited personnel, and on the other, to improve the quality of logistics and, most importantly, to increase road safety.

The article describes an approach to reducing the number of questions in the Keys to Personal Mastery Inventory (KPMI) psychological testing questionnaire with the further addition of missing data and their use in a machine learning model to predict respondent satisfaction with their job. Since the 30s of the 20th century, mathematicians, physicists and engineers have been working on the description of traffic flows. But theoretical models do not work well because a person is driving a car, and a person is unpredictable. In the same situation, it behaves differently. This work aims to reduce driver uncertainty using machine learning methods.

Information security systems that use machine learning algorithms and are based on real datasets are among the most effective. The format and content of datasets used in artificial intelligence and machine learning methods directly affect the quality of information security system implementation. The article presents a classification of datasets in the field of information security of communication networks. The structure and features of several characteristic data sets related to the use of machine learning algorithms in information security systems are shown.

Ensembles of neural networks are used to improve the performance of the model and show themselves well in the case of using deep recurrent and convolutional networks on serial [1] and graphic data [2]. The article discusses the advisability of using a snapshot ensemble of not deep feedforward neural networks on a synthetic dataset.