Важливою складовою лінгвістичного підходу до виявлення аномалій у часових рядах є критерій, за яким оцінюється подібність двох моделей. Наявність аномалії встановлюється шляхом порівняння моделей. Саме від вибору критеріїв залежить можливість застосування лінгвістичного підходу до аналізу часових рядів різної природи. Для вчасної реакції на ситуацію важлива швидкість виявлення аномалії. Вибір критеріїв залежить від можливості застосування мовного підходу до аналізу часових рядів різного характеру. Розглянуті основні метрики схожості текстів Геммінга, Левенштейна, Джаро-Вінклєра та ін.

In this work, we present our research on efficiency of single-core execution of agent-based epidemiological models with contact-tracing transmission.We performed an analysis of existing epidemiological agent-based modeling tools from the perspective of their performance, which reflects computational time needed to perform a single simulation step of a model with fixed number of agents.We developed several simulation algorithms, based on different model types, and showed some optimizations to maximize the performance of them.We designed a metric to compare simulation step execution times on a single CPU core and used to estimate a performance of underlying simulation algorithms in the existing methods with developed one.Even though, as it will be discussed below, it is very difficult to estimate an execution time of some algorithm without actual access to it, we present the results of our method on both modern and old CPU.

В статті розглядається використання гібридного лінгвістичного підходу до моделювання та прогнозування часових рядів. Гібридний лінгвістичний підхід передбачає процес побудови лінгвістичних моделей (ЛМ) з використанням прихованих Марковських моделей (ПММ). Одним з кроків побудови ЛМ є застосування інтервального підходу для розбиття множин значень часового ряду з використанням елементів інтервальної математики та різноманітних розподілів ймовірностей. Будується морфізм перетворення чисельних образів до символьного вигляду, який можна використовувати для вирішення певних задач. Запропоновані методи дають якісні результати на короткостроковому прогнозі, що не відрізняються від прогнозів профільними методами, при використанні менших обчислювальних ресурсів.

In this work, we present our results in development of a domain-specific programming language for an agent-based infectious disease spread modeling with contact tracing transmission. We performed an analysis of methods and tools, accessible to a scientific community, which showed significant problems of currently developed approaches. Some methods provide significant performance, sacrificing ease-of-use, others offer simple programming interface but suffer with performance drawbacks. To address both of these issues, we designed a domain-specific programming language with simple and yet powerful syntax and minimal amount of boilerplate code. The chose of the underlying infectious disease spread model is based on a growing interest in agent-based models, which are becoming more popular due to higher simulation accuracy compared to classical compartmental models. At its core, each agent has its personal user-defined daily schedule, which dictates where the agent is situated at each simulation step based on numerous parameters, such as simulation “day”, agent’s inner state, inner states of linked to this agent compartments, which can be, for example, its house, place of work, school, etc. The underlying language’s engine uses this schedule to check, which agent pair were in contact with each other and where at each simulation step. The infectious disease, defined by a set of rules, which control how do these contacts affect inner states of both agents and how disease progresses in context of a single infected agent. We designed a compiler with Python as intermediate language for a simple programming interface and Numba support for machine language translation. Performed tests showed, that the language is capable of describing complex epidemiological models and the compiler is able to generate machine code efficiently enough to run simulations with medium number of entities. The main problem of a developed approach is the quadratic nature of algorithmic complexity of a contact tracing transmission process. We left optimization of it as a topic for a future research.

This article discusses the implementation of a software library for the analysis of the electrocardiogram signal. A feature of this library is to improve the functionality and simplify the interaction with existing machine learning software and tools for loading, processing and storing ECG signal datasets by using the Word2Vec model. The library increases development speed of a new software, which involves various ECG analysis. Therefore, scientists could more easily implement their ideas related to NLP and ML.

In this paper, we present our findings on applying Markov chain generative model to model actions of an agent in Markov decision process framework. We outlined a problem of current solutions to reinforcement learning problems that utilize the agent-environment framework. This problem arises from the necessity of performing analysis of each environment state (for example for q-value estimation in q-learning and deep q-learning methods), which can be computationally heavy. We propose a simple method of ‘skipping’ intermediate state analysis for which optimal actions are determined from analysis of some previous state and modeled by a Markov chain. We observed a problem of this approach that limits agent’s exploratory behavior by setting Markov chain’s probabilities close to either 0 or 1. It was shown that the proposed solution by $$L^1$$ -normalization of transition probabilities can successfully handle this problem. We tested our approach on a simple environment of k-armed bandit problem and showed that it outperforms commonly used gradient bandit algorithm.

In the article on the stage of intervalization in the construction of a linguistic model of the time series, it is proposed to apply the description of intervals in the form of fuzzy subsets with a probabilistic degree of affiliation. Definitions are considered operations of limiting addition, marginal product and concentration for their hybrid parameters, when one of them is a fuzzy subset in which the degree of belonging is a random variable, and the second parameter is a constant. These operations are used in fuzzy derivation as a direct construction of the division of the area of acceptable values of the time series. The need to use such an approach is dictated by the problem of inverse linguistic transformation. This approach will effectively assess the correctness of the constructed linguistic models of the input time series.

Heart disease accounts for a significant percentage of deaths in both Ukraine and most countries. For example, every year in Ukraine more than 68% of people die from cardiovascular disease. An important factor in the fight against the disease is the prevention and detection of the disease in its early stages. The principal technique of observing the heart is electrocardiography, so it is very important to quickly and accurately analyze the electrocardiogram (ECG). In this article propose to expand the capabilities of automatic analysis of electrocardiograms by creating a Word2Vec model based on selected waves in the ECG.

The paper is dedicated to the overview of current state of the evolutionary games approach to the building of environments to analyze players behavior. The evolutionary game theory applications differ from the orthodox game theory. Initially, it was used to address problems in evolutionary biology and later was suited for broader range of problems.We will oversee the development of the evolutionary games theory in finance and its applications in behavior analysis in competitive gaming. The paper is focused on replicator dynamics, learning model based on it and its possible application to behavior analysis based on fuzzy algorithms and approaches used in economics to be applied to the new emerging field of cybersports.

This work proposes a new approach for identifying heart anomalies on electrocardiograms data using adaptive multistage method of anomalies detection. The method includes: search of exact match pattern with fragments in linguistic chain, analysis by average distance, analysis by fuzzy distance, reduction to a common time grid. Used various matrix of linguistic distances. Described source data and database process filling.

У статті розглядається програмне забезпечення для аналізу і виявлення серцево-судинних захворювань у їх початковий період, щоб забезпечити просте, швидке та набагато діюче лікування пацієнта, а також забезпечити своєчасний контроль хронічних хвороб, що у свою чергу має покращити ступінь охорони здоров’я та зменшити витрати на непотрібні заходи для нього.

На сьогоднішній день виявлення аномалій є однією із головних причин виконання аналізу даних. Із подальшим розвитком інтернету речей, потреба у автоматизованих системах моніторингу та прийняття рішень, здатних вчасно розпізнати збої або помилки в роботі різного роду пристроїв та інфраструктури, та не допустити небажаних наслідків, буде тільки зростати. Саме тому в цій статті дослідження присвячене розробці ефективних алгоритмів виявлення аномалій. Представлені практичні результати аналізу часових рядів цін на акції всесвітньовідомих кампаній.

This work proposes a new approach for identifying heart anomalies on electrocardiograms data using linguistic modeling. The process of identifying anomalies in the proposed approach consists of the following subtasks: the subtask of interval splitting, the subtask of linguistics, the subtask of anomalies searching. The approach includes: the creation of a linguistic pattern database, represent ECG as well as linguistic chain, use linguistic pattern database to search for linguistic chain parts based on abnormal patterns. Linguistic model is suggested for the creation of an anomalies database for further detection in a cardiogram, reproduced in the form of a linguistic chain. Storing ECG as a linguistic model facilitates is easy for data storing and data search in patient history. Linguistic pattern database has filling stages: signal conversion in digital time series, interval splitting, matching interval with an alphabet symbol, creation of alphabetic symbol time series. Anomalies search based on seeking abnormal linguistic patterns in ECG linguistic chains.

A mathematical apparatus which can be used in data analysis problems of different nature for forecasting nonlinear non-stationary processes is expounded.

A problem of construction of matematical models using time series is considered. The concept of model structure is proposed that is used to develop a model construction technique. Criteria are proposed for determining possible nonlinearities with respect to variables. To estimate adequacy of the model constructed, a set of mutually supplementing statistical criteria is given.