ON ONE APPROACH FOR FEATURE SELECTION BASED ON THE APPLICATION OF THE METHOD OF LOGICAL ANALYSIS OF DATA

Предлагается подход для отбора важных признаков при классификации наблюдений. Реализация подхода основана на построении логических правил на базе метода логического анализа данных и учете частоты использования признаков при их формировании для конкретной задачи классификации. An approach is proposed for the selection of important features in the classification of observations. The implementation of the approach is based on the construction of patterns based on the method of logical analysis of data and taking into account the frequency of using features when forming them for a specific classification task.

An iterative feature selection procedure for a classification problem based on the method of logical analysis of data

An iterative procedure for selecting features for classifying observations is proposed. The main principles of the proposed iterative procedure are ranking and selection of features according to the frequency of their use when constructing logical patterns based on the method of logical analysis of data. The empirical confirmation of the expediency of this procedure is given.

REDUCTION OF THE CLASSIFIER IN THE METHOD OF LOGICAL ANALYSIS OF DATA BASED ON THE Ε-, Δ-CRITERION FOR SELECTING PATTERN

Предлагается алгоритмическая процедура редукции классификатора в методе логического анализа данных, основанная на отборе закономерностей с помощью ε-, δ-критерия. Реализация подхода заключается в формировании исходного классификатора как набора закономерностей на базе наблюдений обучающей выборки, применения к полученным правилам процедуры наращивания и последующего их отбора в новый классификатор на базе ε-, δ-критерия. Приводится эмпирическое подтверждение целесообразности данной алгоритмической процедуры. An algorithmic procedure for the reduction of the classifier in the method of logical analysis of data, based on the selection of patterns using the ε-, δ-criterion is proposed. The implementation of the approach consists in the formation of the initial classifier as a set of patterns based on observations of the training sample, application of the increasing procedure to the obtained patterns and their subsequent selection into a new classifier based on the ε-, δ-criterion. An empirical confirmation of the expediency of this algorithmic procedure is given.

LOGICAL ANALYSIS OF DATA WITH PARETO OPTIMAL PATTERNS SEARCH

Предлагается способ генерации логических закономерностей через аппроксимацию множества Парето эвристическим алгоритмом NSGA-II. Указанный метод применяется для решения задачи медицинской диагностики. The paper focuses on logical patterns generation as Pareto set approximation using heuristic algorithm NSGA-II. This method is used to solve the problem of medical diagnostics.

Computational Complexity and ILP Models for Pattern Problems in the Logical Analysis of Data

Logical Analysis of Data is a procedure aimed at identifying relevant features in data sets with both positive and negative samples. The goal is to build Boolean formulas, represented by strings over {0,1,-} called patterns, which can be used to classify new samples as positive or negative. Since a data set can be explained in alternative ways, many computational problems arise related to the choice of a particular set of patterns. In this paper we study the computational complexity of several of these pattern problems (showing that they are, in general, computationally hard) and we propose some integer programming models that appear to be effective. We describe an ILP model for finding the minimum-size set of patterns explaining a given set of samples and another one for the problem of determining whether two sets of patterns are equivalent, i.e., they explain exactly the same samples. We base our first model on a polynomial procedure that computes all patterns compatible with a given set of samples. Computational experiments substantiate the effectiveness of our models on fairly large instances. Finally, we conjecture that the existence of an effective ILP model for finding a minimum-size set of patterns equivalent to a given set of patterns is unlikely, due to the problem being NP-hard and co-NP-hard at the same time.

Reconstruction Rating Model of Sovereign Debt by Logical Analysis of Data

Sovereign debt ratings provided by rating agencies measure the solvency of a country, as gauged by a lender or an investor. It is an indication of the risk involved in investment and should be determined correctly and in a well-timed manner. The current system is lacking transparency of rating criteria and mechanism. The present study reconstructs sovereign debt ratings through logical analysis of data (LAD), which is based on the theory of Boolean functions. It organizes groups of countries according to 20 World Bank-defined variables for the period 2012–2015. The Fitch Rating Agency, one of the three big global rating agencies, is used as a case study. An approximate algorithm was crucial in exploring the rating method, in correcting the agency’s errors, and in determining the estimated rating of otherwise unrated countries. The outcome was a decision tree for each year. Each country was assigned a rating. On average, the algorithm reached almost 98% matched ratings in the training set and was verified by 84% in the test set.

Study of Fitch Rating`s Response to Covid-19 by Logical Analysis of Data

Credit ratings, represent the creditworthiness of countries and financial organizations that nowadays due to the corona virus crisis hit the world is being threatened to be downgraded. This study uses Logical Analysis of Data to analyze the Fitch rating agency response to Covid-19. Three varied parts of variables, composed of the significant economic and social factors, pandemic -related variables and pre-credit rating (2019) are under survey. The time interval of the study is 2019-2020. The output of the research in the form of the decision trees shows the selected patterns of each newly published Fitch rating in July of 2020. The consequences of the research in training and test sets by 100% and 80% matched cases, respectively shed light on the robust results of explored patterns. Surveying on Fitch`s response in this span showed that pandemic-related variables mostly have an impact on “B” classes and they were not significant in “investment grades” (AAA-BBBP), whereas, 2019`s credit rating may be a strong factor to forecast next ratings just in normal state of affairs, nevertheless, selected well-built economic and social factors described the hidden structure of Fitch Agency in the optimum way during pandemic also.