JSM Method Research Articles

The Heuristics of Detection of Empirical Regularities by JSM Reasoning

The heuristics of the detection of empirical regularities by the JSM method of automated support of scientific research is considered. This heuristic consists of applying the JSM reasoning for JSM research, which consist of recognizing the preservation of hypotheses about the causes and hypotheses about predictions that form empirical dependencies for embedded database sequences. Mechanisms for detecting empirical regularities are causal forcing (CF), through which the necessity modalities (for empirical laws), the possibility modality (for empirical tendencies) and the modality of weak possibilities (for suspicious tendencies) are determined. The means of approximate detection of empirical regularities are determined and the results of their application for detecting the possibility of diabetes mellitus in patients with chronic pancreatitis are presented.

TextJSM: Text Sentiment Analysis Method

The TextJSM method of text sentiment analysis is proposed, based on JSM method of automated hypothesis generation. Two versions of the TextJSM method are presented, that is, for solving predictive and descriptive problems. Parallel implementation of the main stages of both versions is considered. Experimental studies based on the ROMIP 2011–2012 seminar text corpora show the superiority of the developed method over other data mining methods.

Causal Inference in Psychological Data in the Case of Aggression

We carried out an empirical study of aggression in relation to different personal traits. In this article we present results obtained for different forms of aggression, including results of machine-learning experiments with the AQJSM method. The method distinguishes several classes with different levels of aggression defined with a special form, as well as making causal inferences with AQ preprocessing and the first stage of JSM method of extraction of cause and effect relationships. The proposed method produces acceptable results for both small datasets and big data with incomplete information.

On JSM reasoning applicable to unions of factbase subsets: Part 1

In this paper, we consider the strengthening of the JSM method for automated support of scientific research (ASSR JSM method) by introducing a ternary relationship of causality such that positive facts can contain negative causes and negative facts can contain positive causes. In this connection, we consider unions of factbase subsets that correspond to combinations of pairs of similarity predicates that, in turn, represent binary and ternary relationships of causality, respectively, for formalization of JSM reasoning rules.

ИЗУЧЕНИЕ СОЦИАЛЬНЫХ ПРЕДСТАВЛЕНИЙ С ИСПОЛЬЗОВАНИЕМ МЕТОДА ВИНЬЕТОК: КОЛИЧЕСТВЕННЫЙ ПОДХОД

The article focuses on the prospects of creating vignettes as a new method in empirical sociology. It is a good alternative to the conventional mass survey methods. The article consists of a few sections differing by the focus. The vignette method is not popular among Russian scientists, but has a big history abroad. A wide range of problems can be solved by this method (e.g. the prospects for guardianship and its evaluation, international students’ adaptation to the educational system, social justice studies, market-ing and business research, etc.). The vignette method can be used for studying different problems including sensitive questions (e.g. HIV, drugs, psychological trauma, etc.), because it is one of the projective techniques. Projective techniques allow to obtain more reliable information, because the respondent projects one situation on the another, but at the same time responds to his own stimulus. The article considers advantages and disadvantages of the method. The authors provide information about the limitations of the method. The article presents the key principles for designing and developing the vignettes method depending on the research type. The authors provide examples of their own vignettes tested in the course of their own empirical research. The authors highlight the advantages of the logical-combinatorial approaches (especially the JSM method with its dichotomy) for the analysis of data in quantitative research. Also they consider another method of the analysis of the data that implies the technique of “steeping”, i.e. when the respondent gets new information step by step, which extends his previous knowledge.

Возможности ДСМ-метода для построения социологических гипотез (на примере анализа политического участия)

This article demonstrates the possibilities of the method of formalized qualitative analysis of sociological data (FQASD) for creating sociological hypotheses. FQASD is a version of the JSM Method which is tailored to solve sociological problems. The JSM Method is a logical-combinatorial (non-statistical) method based on using actual logic tools and designed for analyzing causal relations between feature combinations (object characteristics) and the sought-for effect or a variety of effects (object properties).The basic ideas underlying the philosophy of this method include the principles of empirical induction, structural analogy and abduction (in the interpretation of Charles Peirce) as a means of accepting hypotheses. The authors provide theoretical foundation and a brief description of the procedures they have developed.The main investigative question is — who are party activists in medium-sized Russian cities? An additional question — what groups of people can be considered the closest to political activists based on civic participation criteria?To answer these questions we used the FQASD method. First of all, political participation is shown by a small fraction of the voting public (which makes it impossible to use statistics).Second — we are dealing with a great number of characteristics which presumably determine the sought-for behavior (political participation) and we assume that they might have a complex effect.In addition to political activists, we’ve constructed three more groups: Civic Activists, Individual Helpers and Inactive People. The characteristics of political activists were consistently compared to the characteristics of these three groups. In such a comparison they became antipodes to each other. The most non-trivial and substantial conclusion we arrived to concerns how respondents define their own subjectivity and the subjectivity of their fellow citizens. Political activists consistently attribute social subject status to themselves and deny this status to others. In this respect they correlate with the proportion of the respondents with the weakest resource base. Civic activists, being antipodes of political activists, recognize the social subject status of their fellow citizens.The advantages of the FQASD in this and similar cases are that: 1) the results obtained were not formulated as preliminary hypotheses; 2) the results give new directions for analyzing the social activity of ordinary citizens; 3) the method demonstrated the possibility of revealing sociological patterns in small samples.

ДСМ-метод в качественном социологическом исследовании: основные принципы и опыт использования

This article outlines the ideology and the basic principles of the JSM Method created by V.K. Finn and his colleagues. This method already has an extensive application background when it comes to solving sociological problems. The first experiments were carried out in 1997–1998, and since then this method’s possibilities and limitations have been defined, important procedural solutions have been found which allow sociologists to explore sociological patterns based on formalized analysis of various sorts of empirical data (materials surveys, statistics, and observations). The article shows how the JSM Method in its version for solving sociological problems (formalized qualitative analysis of sociological data — FQASD) differs compared to qualitative comparative analysis, which was developed by American sociologist C. Ragin.Methodical solutions are illustrated by examples of empirical studies conducted utilizing this method. One of the tasks was situation analysis (analyzing the context of the examined process). It was fulfilled based on materials from a study of a certain integral feature of labor behavior (namely — employee “stability-loyalty”). Certain parameters which describe the labor market in the city have shown what said feature depends on. The second task was analyzing the internal structure of a certain integrated sociological feature, examining the structure of the “job satisfaction” concept as an example. A certain empirical regularity has been revealed using the JSM Method: job satisfaction may include dissatisfaction with the majority (and sometimes all) of the labor situation parameters. A conclusion is made on the possibilities of using the JSM Method in empirical sociological research. In particular, the JSM-method helps verify the validity of the meaning of those generalization concepts used by the analyst, as well as permit the inclusion of varying types of data into analysis (including non-numeric), and, finally, identify behavior determinants.

J.S. Mill’s similarity method and difference method in the language of category theory

This article describes the application of the language of category theory to comparison with a sample in order to formalize the similarity method and the difference method, which underlie the JSM method for automatic hypotheses generation.

A technique for retrieving cause-and-effect relationships from optimized fact bases

A technique for retrieving causal connections of binary relationships from a set of fact bases is suggested. The fact bases are formed for the target properties of each class of objects. The class descriptions are formed by habituation on data from a loosely formalized object domain. The habituation is organized using a co-evolutional genetic algorithm that reduces the initial feature space. The cause-and-effect relationships for all target properties are sought by the formed optimized class descriptions using stage one of the JSM method. The suggested technique is suitable for the analysis of the full data in small amounts and large incomplete data arrays. Several model experiments using the MIMIC II medical database were conducted.

An approach to text data categorization based on the ideas of J.S. Mill

The problem of the automatic categorization of text documents in the natural language is considered. The categorization is made by a method that is based on ideas of J.S. Mill. This technique uses the general principles (but not the technical details) of the JSM method for the automatic generation of hypotheses. Tests are described and the performance quality of the system that was built to carry out the described technique is assessed. With an optimal selection of options, the suggested approach shows better accuracy than other techniques.

WEBCHEM-JSM: an information environment for implementation of the JSM method in pharmacology

WebChem-JSM is a software system that is an information environment for application of the JSM method in the discovery of structure---activity relationships in pharmacology and related fields. A client---server architecture based on web technology is proposed to solve tasks. Compared to past solutions of this problem, centralized storage of data and knowledge together with the ability to accumulate knowledge are attained. The modular architecture of the system provides rich capabilities for deploying various versions of the components, for comparative study of solvers and strategies of JSM method. For the first time, a parallel JSM solver is used in an intellectual JSM system to achieve greater performance.

On the sufficiency of the grounds for accepting the results of data mining using the JSM method

The problem of the control of the sufficiency of the grounds for accepting the results of the generation of empirical relationships from data via the JSM method is discussed. A procedural structure based on the analysis of the system of equivalence classes covering corresponding similarity classes is formulated in order to implement such control. Relationships between the proposed control mechanism of the sufficiency of grounds with basic structures of the JSM method are presented (rules of plausible inference of the first and second kind).

Extraction of cause-effect relationships from psychological test data using logical methods

A shared use of the AQ learning and JSM method in extracting cause-effect relationships from psychological test data is considered. The AQ learning is used to descript a test group using rules. The group description is a basis for constructing the factbase of the JSM method. The first stage of the JSM method is used for hypothesizing on cause-effect relationships in the factbase. A complete algorithm of the developed method is described and recommendations for its use are provided. A contrasting analysis of the results of the proposed method and statistical processing results is performed.

On the intelligence of the intelligent data analysis

Some functional features of software tools for computer data analysis are described, which make it possible to assign software tools of this type to the class of intelligent systems (in the meaning of this term that is traditional for research in artificial intelligence). The abilities of intelligent data analysis (IDA) are demonstrated based on the example of the formalization and solution of the problem of prediction of resource intensity for projects of design and the production of complex software and hardware systems for HVAC equipment for the telecommunications industry using the JSM method for automated hypothesis generation. The basic descriptive and argumentation capabilities of the JSM method are demonstrated, which make it possible to use it as an IDA platform.

Epistemological foundations of the JSM method for automatic hypothesis generation

This article consists of two parts. Part I considers the so-called problem of induction, viz., the history of its genesis and elaboration, and formulates the main principles and logical tools of the JSM method for automatic hypothesis generation, including JSM reasoning. JSM reasoning is demonstrated to be a synthesis of three cognitive procedures: induction, analogy, and abduction. Two possible kinds of abduction are considered. Part I focuses on formal tools to be used for the solution of the natural-scientific problem of induction. Part II formulates the procedural semantics for predicates that are used to formalize JSM reasoning, considers quasi-axiomatic (open) theories used to represent knowledge in intelligent systems, determines the means for assessing the quality of JSM-reasoning results, and, finally, suggests a solution of the naturalscientific problem of induction for intelligent systems. Part I was published in the journal Nauchno-Tekhnicheskaya Informatsiya, ser. 2, 2013, no. 9, pp. 1---29. Part II was published in the journal Nauchno-Tekhnicheskaya Informatsiya, ser. 2, 2013, no. 12, pp. 1---26.

Application of Case-based Methodology for Early Diagnosis of Computer Attacks

In this article we consider the mathematical foundations and software implementation of the early diagnosis of computer attacks. For this we used the JSM method of automatic hypothesis generation and the theory of case-based models. This software outputs hypotheses about the properties and expected consequences of a new computer attack. The system analyses a set of properties of the computer attack known to the user. For this we use the Base of the cyber attack’s precedents, described in the language of fuzzification of Boolean-valued models. Each potential property of the new attack is studied by using the JSM method. This process builds sets of positive and negative hypotheses concerning each property, giving a set of properties and consequences characteristic of the attack that has yet not happened at the time of analysis. The developed algorithm has polynomial complexity.

Feature selection using decision trees in the problem of JSM classification

A method of feature selection is considered for training a classification algorithm based on the JSM method. The method is based on the use of decision trees. It includes construction of a maximal tree using the C4.5 algorithm and the preparation of a series of truncated trees based on the criterion of minimal cost complexity. The results and parameters of the JSM classifier with different numbers of selected features are compared.

Comparing discretization and selection methods for the logical-combinatorial classification of continuous parameters

Preprocessing methods for handling problems with features containing continuous attributes are discussed for learning a classification algorithm based on the JSM method. Discretization methods for continuous parameters that do not make use of class information on feature distribution are compared to entropy-based methods employing class labels in interval partitioning. An entropy-information-based method for selecting attributes is also discussed.

On the functionality of the causal relationship used in JSM reasoning

The property of the functionality of the causal relationship (a subobject is the cause of multiple properties) used in the JSM method for the automatic generation of hypotheses is considered. The functionality of this relationship on each fixed step of the JSM reasoning is shown. Some features of the studied characteristics of the causal relationship on sequences of JSM reasoning steps are examined.

J.S. Mill’s inductive methods in artificial intelligence systems. Part II

In the second part of this article we define the relationship of derivability of hypotheses from fact bases, which is used to detect different types of correctness of JSM reasonings. A correspondence is established between the total correctness of JSM reasonings and tolerance spaces. Inductive methods of residues and concomitant variations are formalized and the appropriate strategies of the JSM reasoning are defined. Using the rules of plausible inference for the method of concomitant variations operational definitions of dynamic patterns in fact bases. Features of the JSM method for the automatic generation of hypotheses in intelligent systems as a means of knowledge discovery are also discussed.