Якуба Вячеслав Иванович

Short-Term Covid-19 Incidence Prediction in Countries Using Clustering and Regression Analysis

2023 · CHAPTER · en

As of February 07, 2022, more than 395 million cases of COVID-19 had been identified in the world, with 5.74 million deaths. The paper considers methodology for predicting the number of cases in the short term using a preliminary assessment of countries based on three indicators: expert assessments of the law-abiding population, the level of education and restrictive measures taken in the country. The description and composition of the groups obtained are given. An assessment of the accuracy of the forecast results is made. A comparison of the considered models of 2020 with 2022 is given.

Network analysis of publications on studies of Parkinson Disease

2023 · ARTICLE · en

More than 10 million people worldwide are suffering from Parkinson Disease (PD). We analyse publications on studies of PD over the period from 2015 to 2021. We have collected about 70 thousand papers from 4912 different journals. After the data preprocessing 39811 publications and 3292 journals are left. Methods of centrality analysis have been applied to identify the most significant publications and journals in the field. Classical and new centrality indices have been evaluated for the whole period and for each year. It can be noticed that changing the parameters makes it possible to extract groups of specialized journals and fields of study, which intensively cite each other.

On the Individual and Coalitional Manipulability of q-Paretian Social Choice Rules

2023 · CHAPTER · en

We study the degree of individual and coalitional manipulability of q-Paretian social choice rules under Impartial Culture. Manipulability is defined as a situation, when an agent or a coalition, which consists of some agents, misrepresents her/their preferences to obtain a better outcome of the social choice rule. We study a class of q-Paretian social choice rules, which consists of four rules: Strong q-Paretian simple majority rule, Strong q-Paretian plurality rule, Strongest q-Paretian simple majority rule, and Condorcet practical rule. For the cases of 3, 4, and 5 alternatives and for the cases from 3 to 100 agents, we use computer modelling to calculate a number of manipulability indices. We provide the analysis of the results for different cases.

Superposition Principle for Tornado Prediction

2023 · CHAPTER · en

Further Results on the Manipulability of Social Choice Rules—A Comparison of Standard and Favardin–Lepelley Types of Individual Manipulation

2021 · CHAPTER · en

Экономические журналы России, их характеристики и сетевой анализ

2021 · ARTICLE · ru

В статье изучаются экономические журналы России, входящие в базу Российского индекса научного цитирования (РИНЦ). Исследуется цитируемость 466 журналов за период 2016–2020 гг. и сопоставляются полученные рейтинги по таким характеристикам, как совокупная цитируемость, индекс Хирша, двухлетний импакт-фактор РИНЦ и Science index РИНЦ. Также проводится сетевой анализ цитируемости с использованием сетевых индексов центральности, которые определяют влиятельность журналов: индекс Коупленда и новые индексы Bundle и Pivotal. Единственный журнал, который во всех категориях входит в топ-5, — это «Вопросы экономики». Другие журналы в списке лидеров отличаются в зависимости от используемой метрики. Стоит отметить, что по ряду показателей в лидеры выбиваются издания, к репутации которых есть вопросы у научного сообщества. Основной причиной является большой объем публикаций в подобных журналах, который делает классические журналы менее значимыми с точки зрения индексов.

Constructing an Efficient Machine Learning Model for Tornado Prediction

2020 · ARTICLE · en

Tornado prediction variables are analyzed using machine learning and decision analysis techniques. A model based on several choice procedures and the superposition principle is applied for different methods of data analysis. The constructed model has been tested on a database of tornadic events. It is shown that the tornado prediction model developed herein is more efficient than a previous set of machine learning models, opening the way to more accurate decisions.

Power Distribution in the Networks of Terrorist Groups: 2001–2018

2020 · ARTICLE · en

Since 9/11, terrorism has become a global issue of the twenty-first century. Terrorist organizations become important actors of world politics as they gain influence on political process and decision-making. Some organizations compete with each other in order to gain more power and influence. We study the distribution of power among terrorist groups using network approach and applying classic and new centrality indices (Short-Range (SRIC) and Long-Range interactions indices (LRIC)). These indices allow to identify terrorist groups with direct and indirect influence on the terrorist network.

Matrix-vector approach to construct generalized centrality indices in networks

2020 · PREPRINT · en

Вводится новый класс индексов центральности в сетях, в которых учитываются параметры вершин, прямые и косвенные (с фиксированной длиной пути) связи между вершинами, а также влияние групп вершин на вершину. Полное влияние оценивается в виде линейной комбинации его компонент, но могут использоваться и другие правила.

Manipulability of Majoritarian Procedures in Two-Dimensional Downsian Model

2020 · CHAPTER · en

For the two-dimensional Downsian model the degree of manipulability of 16 known aggregation procedures, based on the majority relation, is evaluated using the Nitzan-Kelly index. Extended preferences for multi-valued choices are used to evaluate the fact of manipulation. Individual manipulability of agents is considered, when manipulating agent moves its ideal point over the plane. The range of possible manipulating positions of the agents is restricted to some rectangle on the two-dimensional coordinate space, within the feasible area of positions of alternatives and agents. The preferences of agents are assumed to be linear orders, constructed by the proximity of the alternatives to the agents, ordered according to Euclidean distance. The computer calculations, using Monte-Carlo simulations has been performed for 3, 4, and 5 alternatives and for even number of agents from 4 to 20. 100 thousands profiles were generated for each number of alternatives – number of agents case. The results of the simulations show that there are groups of procedures with relatively low degree of manipulability for all of the considered multiple-choice extensions.