Video game tutorials are crucial because they shape a player's initial experience with a game. The design of these tutorials poses a significant challenge as they can influence the continuous engagement of players. Existing tutorials often employ a one-size-fits-all approach, failing to account for individual skill variations among players. This paper introduces GameMentor, a creative Al-driven tutorial system that personalizes the learning experience for each player. GameMentor employs a deep reinforcement learning AI agent trained in gameplay to analyse gameplay data and identify player's mistakes and behaviour at critical moments. This information is then used to create customized tutorials. They allow players to practice by recreating the scenarios where they initially struggled, thereby enhancing their learning pace and overall performance. The results of our study demonstrate significant benefits of the GameMentor compared with traditional tutorials. GameMentor offers a more tailored and effective learning experience.

Easy access to large amounts of data triggers an interest in various data analysis techniques. On multiple occasions, an insightful and meaningful ‘view’ of a phenomenon represented by the data is not easy to spot in the obtained results. The users would like to learn more about essential aspects in the context of their own perception of considered concepts and gain a more human-like and linguistic-based interpretation of the results. Yet, existing techniques cannot work with imprecise linguistic terms and do not provide users with results expressed in a human-friendly format.This paper proposes a simple data analysis technique based on calculating correlation. The method investigates the relationships between linguistic terms defined – individually by the users – on data features. It leads to personalized summarization of data that includes the essential relations. In addition, we use the proposed method on the temporal facet/angle of data allowing us to provide the users with the inside regarding the temporal volatility of relations.

The increased popularity of Linked Open Data (LOD) and advances in Natural Language Processing techniques have led to the development of Question Answering Systems (QASs) that utilize Knowledge Graphs as data sources. QASs perform well on simple questions providing precise and concise answers. Yet, most of them cannot process answers that contain a large volume of numerical values and are not able to provide users with answers in a human-friendly format. In this paper, we propose a user-defined method for constructing linguistic summarization of multi-feature data. It selects suitable summarizers and quantifiers and works with linguistic constraints imposed on the data. The method relies on definitions of linguistic terms constructed by users using an easy and simple graphical interface. Additionally, we introduce a Context-based User-defined Weighted Averaging (CUWA) operator. It allows determining an average value of data that satisfies multiple constraints that are account for the context defined by the user. We include several illustrative examples.

Networks have a rich history of capturing interactions and are widely used in fields ranging from medicine to sociology to computer science. A network where links between nodes may exist on different relations - called a multilayer network - is a flexible and increasingly common type of network since it allows for describing more complex interactions between nodes. Modeling such networks is thus an important task as it allows machines to reason with them intelligently and perform tasks such a link prediction, node classification, and community detection. Two common approaches to modeling multilayer networks are blockmodeling and embedding models. In this paper, we propose a model which fuses these approaches in a neural architecture, allowing it to leverage their respective strengths. To do this, we follow the standard blockmodeling approach of assigning nodes to communities and extending community interactions to node interactions. Community assignment is inspired by embedding models such that each node has an embedding which generates its community assignment. We account for network multilayeredness by mixing a relation embedding with the community interactions vector which acts as a relational modifier on community interactions. We evaluate our model on the three aforementioned tasks using real-world datasets and compare against state-of-the-art models. We find that performance by state-of-the-art models is highly dataset dependent and that our model achieves comparable results. We conclude that blockmodels may be successfully fused with embedding models for network modeling using a neural architecture and that simply mixing a relation embedding with community interactions is a viable way of handling for network multilayeredness.

The Mixed-Membership Stochastic Blockmodel (MMSB) is proposed as one of the state-of-the-art Bayesian relational methods suitable for learning the complex hidden structure underlying the network data. However, the current formulation of MMSB suffers from the following two issues: (1), the prior information (e.g. entities' community structural information) can not be well embedded in the modelling; (2), community evolution can not be well described in the literature. Therefore, we propose a non-parametric fragmentation coagulation based Mixed Membership Stochastic Blockmodel (fcMMSB). Our model performs entity-based clustering to capture the community information for entities and linkage-based clustering to derive the group information for links simultaneously. Besides, the proposed model infers the network structure and models community evolution, manifested by appearances and disappearances of communities, using the discrete fragmentation coagulation process (DFCP). By integrating the community structure with the group compatibility matrix we derive a generalized version of MMSB. An efficient Gibbs sampling scheme with Polya Gamma (PG) approach is implemented for posterior inference. We validate our model on synthetic and real world data.

Resource Description Framework (RDF) is an important way of representing data on the Web. Although RDF is a data format suitable for publishing individual pieces of information together with relations between them, it represents a challenging format for answering questions. Thus, a system and a user interface that are easy and intuitive for users to access and operate on RDF data are of significant importance.In this paper, we introduce a Question-Answering (QA) system that allows users to ask questions in English. The uniqueness of this system is its ability to answer questions containing linguistic terms, i.e., concepts such as SMALL, LARGE, or TALL. Those concepts are defined via membership functions drawn by users using a dedicated software designed for entering `shapes' of these functions. The system is built based on an analogical problem solving approach, and is suitable for providing users with comprehensive answers. We demonstrate the capability of the proposed QA system by answering questions asked over two RDF stores: DBpedia and Wikidata.

The currently observed developments in Artificial Intelligence (AI) and its influence on different types of industries mean that human-robot cooperation is of special importance. Various types of robots have been applied to the so-called field of Edutainment, i.e., the field that combines education with entertainment. This paper introduces a novel fuzzy-based system for a human-robot cooperative Edutainment. This co-learning system includes a brain-computer interface (BCI) ontology model and a Fuzzy Markup Language (FML)-based Reinforcement Learning Agent (FRL-Agent). The proposed FRL-Agent is composed of (1) a human learning agent, (2) a robotic teaching agent, (3) a Bayesian estimation agent, (4) a robotic BCI agent, (5) a fuzzy machine learning agent, and (6) a fuzzy BCI ontology. In order to verify the effectiveness of the proposed system, the FRL-Agent is used as a robot teacher in a number of elementary schools, junior high schools, and at a university to allow robot teachers and students to learn together in the classroom. The participated students use handheld devices to indirectly or directly interact with the robot teachers to learn English. Additionally, a number of university students wear a commercial EEG device with eight electrode channels to learn English and listen to music. In the experiments, the robotic BCI agent analyzes the collected signals from the EEG device and transforms them into five physiological indices when the students are learning or listening. The Bayesian estimation agent and fuzzy machine learning agent optimize the parameters of the FRL agent and store them in the fuzzy BCI ontology. The experimental results show that the robot teachers motivate students to learn and stimulate their progress. The fuzzy machine learning agent is able to predict the five physiological indices based on the eight-channel EEG data and the trained model. In addition, we also train the model to predict the other students’ feelings based on the analyzed physiological indices and labeled feelings. The FRL agent is able to provide personalized learning content based on the developed human and robot cooperative edutainment approaches. To our knowledge, the FRL agent has not applied to the teaching fields such as elementary schools before and it opens up a promising new line of research in human and robot co-learning. In the future, we hope the FRL agent will solve such an existing problem in the classroom that the high-performing students feel the learning contents are too simple to motivate their learning or the low-performing students are unable to keep up with the learning progress to choose to give up learning.

Analysis of large amounts of data requires considerable efforts in order to extract a useful information or build a model of phenomena represented by the data. On multiple occasions data processing activities can be simpler if some understanding of data is obtained. Further, such activities could be even abandoned when a good insight into data is gained. In this paper, we propose and describe a user-friendly approach for examining datasets via their summarization. It is accomplished using questions equipped with linguistic terms representing imprecise yet easy to grasp concepts by humans. Changes in definitions of the terms done by a user result in different answers and allow to learn more about the data. The approach involves an iPad application that provides a means for an easy way to enter definitions of linguistic terms representing user's perception of imprecise concepts.

Social networks are becoming increasingly important in many fields, from marketing analysis to bioinformatics. Link prediction processes are essential tasks required for analysis of the networks' structures. In this paper, we propose a fuzzy computational model, called Fuzzy Social Signature, to represent a network from the perspective of a single user. This model assumes that not all links are equally important and that the relationships between nodes of a social network can be vague and uncertain. Based on the proposed Fuzzy Social Signature, a preliminary technique for link prediction between users performing same activities is proposed. Encouraging results have been obtained with an initial set of experiments using a real-world dataset.

Latent community models are successful at statistically modeling network data by assigning network entities to communities and modelling entity relations as the relations of their communities. In this paper, we describe the limitation of these models in inferring relations between two communities when the entity relations between these communities are unobserved. We propose a solution to this problem by factorizing the community relations matrix into two community feature matrices, thereby adding a dependency between community relations. We introduce the deep dynamic mixed membership stochastic blockmodel based network (DDBN) to demonstrate the feasibility of such an approach. Our model marries the mixed membership stochastic blockmodel (MMSB) with deep neural networks for rich feature extraction and introduces a temporal dependency in latent features using a long short-term memory unit for dynamic network modeling. We evaluate our model on the link prediction task in static and dynamic networks and find that our model achieves comparable results with state-of-the-art methods.

There is no doubt that Internet of Things becomes an important component of future governmental, industrial, commercial and private infrastructures. Interconnected devices, from intelligent ones to simple sensors, will continuously generate enormous amounts of data. It seems impossible to have all this data being transmitted to dedicated processing centres. More and more often, attention is being put on different forms of local, multi-source and multi-level data processing schemas. In this paper, we propose a novel approach to process sensory data in a multi-level fashion. We use elements of Evidence Theory and adopt a newly developed method suitable for satisfying uncertain targets to assess the most adequate state of monitored system/phenomena. We perform this in stages, where observable values are required at the lowest level of processing, while calculations occurring on higher levels use the results of lower level computations. At the same time, levels of belief in the assessed states of a system/phenomena are determined.

Introduction of graph-based data representation formats, that resulted in Knowledge Graphs and Linked Open Data, enables new ways of processing and analyzing relations between individual pieces of data. One of the most important features of such representation is its ability to represent data semantics. We state that an important step towards obtaining a full utilization of graph-based semantics is to create a formal process of extracting underlying structures of data from Knowledge Graphs and Linked Open Data, as well as building data models. The paper proposes a methodology, based on category theory, for representing graph-based data as a topos category. Construction of topos give us the ability to identify two types of features: ones that are involved in definitions of other concepts; and ones that show how other concepts are involved in a definition of a given concept. Topos and structures of features allow for reasoning about concepts and their interrelations. Further, mechanisms of category theory enable to synthesize new concepts. A simple example is included.

A search process becomes an essential component of everyday routine for many users. Users constantly look for new, and more or less relevant items that they require for work or for entertainment. On multiple occasions, they try to find other users who match their `likes' and `dislikes'. Many different methods and approaches have been proposed and developed to address such needs. The Pythagorean Fuzzy Sets have been proposed as a new class of non-standard fuzzy sets. They are related to the idea of Pythagorean membership grades (a, b) that satisfy the requirement a2+ b2≤ 1. The interesting aspect of those types of sets is their ability to express a positive support a - a positive membership grade, and a negative support b - a negative membership grade. In this paper, we propose a method based on the application of Pythagorean fuzzy relations for identifying a degree of matching between users based on their evaluations of items. We use triangular compositions to determine users that match positive evaluations, and users that agree on negative ones. The usage of Pythagorean fuzzy sets allows us to take into consideration both positive and negative aspects of evaluations and find users who like or dislike at least the same items as a given user likes or dislikes. The proposed approach is used to identify users that evaluate movies in a similar way.

Most image segmentation methods based on active contour model are sensitive to the contour initialization. For the images of complex contents, it is difficult to initialize the contour properly and the biased initialization may lead to low-quality segmentation results. Aiming to tackle this problem, we propose an ensemble strategy to improve the contour-based segmentation. The optimal segmentation ensemble is obtained through maximizing the weighted mutual information between the probability distributions of multiple segmentation results. Experimental results validate that the ensemble of contour-based segmentation is robust to the biased initialization and produces stable and precise results for the images of complex contents.

Design and development of complex systems require collaborative efforts of multiple individuals. Availability and usage of formal approaches to specify designed components is essential for their successful completion and application. Usage of specification languages addresses such needs. In the domain of fuzzy systems an XML-based specification language called Fuzzy Markup Language (FML) has been recently standardized. This paper illustrates benefits of its application. The emphasis is put on a process of automatic validation of model specifications, as well as importance of XML schema that defines components of fuzzy systems and constrains that are imposed on them. The validation process that ensues soundness and completeness of specified fuzzy models leads to a simple and straightforward implementation of these models. The paper shows important aspects of such an approach.

We present a methodology to support urban planners and decision makers in obtaining a good awareness of how city assets (points of interest) are perceived by a community, and on the impact and influence that this collective perception can have on other city assets and city issues such as mobility, environment, security. The methodology employees Fuzzy Cognitive Maps and Fuzzy sets. Fuzzy Cognitive Maps are used to model the relationships between elements of mental representations that different communities have with regards to city issues. The concept of signature as relation between two fuzzy sets is adopted, in analogy to what proposed by Yager and Reformat [1], to characterize a point of interest. Different signatures are subsequently grouped to characterize an area and adopted, in combination with sentiment analysis, to derive a measure of collective perception on the quality of the area. This measure is used to activate some qualitative concept of a Fuzzy Cognitive Map and perform what-if analysis. The methodology has been applied to a sample of three POIs (representing three attractions of the city of Salerno) by using data gathered from the Web and involving some real citizens. Our preliminary results are encouraging with regards to the possibilities offered by our approach of enforcing city decision makers with a good awareness on how changes in the perception of quality of urban areas can influence other city related issues.

The Semantic Web format of data representation called Resource Description Framework (RDF) is used in Linked Open Data to represent and store any type of information on the Internet. Its intrinsic feature of being a graph-like format provides opportunities for new approaches of analyzing and absorbing information. The extended format of participatory learning based on propositions as pieces of information is an attractive way of integrating new knowledge. Further, it has been recognized that information can have a disjunctive or conjunctive form. This paper uses the RDF-based participatory learning process to absorb conjunctive information characterized by different degrees of consistency. The modification of participatory learning process that allows for integrating such information is presented, and a simple case study is included.

We present an application of Granular Computing and Three-way decisions to intelligence analysis. In particular we extend the Analysis of Competing Hypotheses with an additional perspective devoted to support analysts in reasoning with groups of hypotheses that can be equivalent on the basis of partial and incomplete evidence, and in classifying these groups of hypotheses with respect to a decisional attribute of interest for the analyst, such as dangerous or safe. Creating and reasoning with granules and multi-level granular structures give to our approach an added value when dealing with a large number of evidence and hypotheses. Three-way decision making offers the possibility of a rapid understanding of how granules of hypotheses approximate a class of dangerous hypotheses, with clear benefits when analysts have to take decision on classifying a group of hypotheses or setting a proper level of attention to group of equivalent hypotheses.

Linked Open Data (LOD) consists of numerous data stores that are highly interconnected. LOD stores use Resource Description Framework (RDF) as a data representation format. A graph-based nature of RDF brings an opportunity to develop new approaches for accumulating data from multiple sources characterized by different levels of confidence in them. Recently, a participatory learning mechanism has been extended to cope with RDF. It is an attractive way of integrating new pieces of information with already known ones. Further, it has been recognized that pieces of information describing entities can have a disjunctive or conjunctive form. This paper uses an RDF-based participatory learning process to aggregate information obtained from multiple data stores. This process provides mechanisms that determine overall certainty in combined data based on levels of confidence in already known pieces of information and new ones. The behavior of such a process used for integrating information equipped with different levels of uncertainty is presented, and a simple case study is included.

The introduction of Fuzzy Markup Language (FML) in 2004 has initiated an important trend in Computational Intelligence research: the application of new web technologies to create more flexible and hardware independent environment for deploying "fuzzy ideas". FML allows researchers and engineers to focus on problem solving activities bypassing additional difficulties related to programming or physical equipment constraints. From that moment on, many researches have been using FML and other XML-based languages for modeling and developing fuzzy systems. However, in spite of their hardware interoperability, XML languages are able to model a fuzzy system in static way and, consequently, they do not provide any support for modelling "evolving" and temporal-based fuzzy systems, as such Timed Automata based Fuzzy Controllers. To address this deficiency, this paper introduces an extension of FML called FMLScript. It is based on a scripting language concept and allows for modeling XML-based systems that can dynamically modify their configurations. As the consequence, a better expressive power can be achieved when compared with static modelling approaches. This is shown in a case study involving a smart grid control.

People recognize familiar faces in a similar way by using interior facial features (facial regions) such as eyes, nose, mouth, etc. However, the importance of these regions in the realization of face identification and a quantification of the impact of such regions on the recognition process could vary from one region to another. An intuitively appealing observation is that of monotonicity: the more regions are taken into account in the recognition process, the better. From a formal point of view, the relevance of the facial regions and an aggregation of these pieces of experimental evidence can be described in the formal setting of fuzzy measures. Fuzzy measures are of particular interest with this regard given their monotonicity property (which stands in a clear contrast with the more restrictive additivity property inherent to probability–like measures). In this study, we concentrate on the construction of fuzzy measures (more specifically, $$ \lambda $$ λ -fuzzy measure) and characterize their performance in the problem of face recognition using a collection of experimental data.

Fuzzy ontologies have been employed to represent and reason over fuzzy information, which often occurs in real-world applications. Fuzzy inference systems (FIS) are well-known computational intelligence systems whose inferences can also be exploited in fuzzy ontology-based applications. Specifically, the combination of fuzzy ontologies and Mamdani-type FIS can provide inferences involving fuzzy rules and numerical property values, which can be considered in other fuzzy ontology reasoning tasks. In this sense, this paper proposes a hybrid reasoner combining fuzzy ontology and Mamdani inference to provide meaningful inferences that are not available to fuzzy ontology-based applications in an integrated way. Fuzzy rules are represented with Fuzzy Markup Language, providing an abstraction level with regard to the underlying FIS implementation. Some experiments are presented regarding a recommender system context, including a comparison with a fuzzy description logic reasoner in terms of fuzzy rule reasoning semantics and integration issues.

The amount of information stored on the Internet is constantly growing, users become more involved in Web activities, and the Web utilization is on the rise.The Internet becomes a place for variety of actions and events related to work, education, and entertainment.The users become aware of the enormous amount of data and information available to them, and in the case of tools supporting utilization of such repository their needs and expectations are growing.Search engines and recommender systems-although helpful-provide limited assistance.The users implore for more-they are looking for tools that filter out irrelevant information, search through nonstructural data, and provide suitable answers to their questions, all in a personalized way and according to their needs.Furthermore, users want to interact with the Web in a more natural way.The soft computing technologies developed around Zadeh's concept of fuzzy sets provide a framework suitable for processing and presenting data in a more human-like way.A significant amount of research has been conducted in the applications of soft computing to analysis and processing of data, modeling, decision making, and knowledge extraction.Maturity of these methods and the growing importance of the Web create a setting suitable for applying soft computing technologies to develop methods and techniques addressing the users' needs and concerns mentioned above.This special issue gives the readers a glimpse how soft computing techniques can be applied to processing data in the environment of the Web.All contributions are extended versions of the papers presented during the First World Conference on Soft Computing, May 23-26, 2011, San Francisco, CA.The paper entitled "A Fuzzy Ontology for Database Querying with Bipolar Preferences" by Nouredine Tamani, Ludovic Lietard, and Daniel Rocacher focuses on obtaining relevant information from data stored in distributed and heterogeneous databases.Queries to the databases result in responses containing relevant as well as irrelevant information.To separate the responses, so the user is provided only with the most relevant data, the paper introduces a semantic-based personalized data access method.The proposed method is governed by evaluating user preferences represented as fuzzy bipolar conditions, i.e. conditions containing both negative and positive declarations.The paper describes a new approach for querying complex information systems that combines a reasoning mechanism (an ontology based on

The web is perceived as a new social platform. Very often, the users look at the web as a place where they can find an individual or group of people with the same or similar interests, or even find new friends. Such situation is reflected in one of the aspects of the web 2.0 called tagging. Tagging is a process of labeling (annotating) digital items-resources-by users. The labels-tags-assigned to those resources reflect users' ways of seeing, categorizing, and perceiving particular items. In general, a single user can label a number of items with a number of different tags. The results of this activity-labeled items and used tags-can be perceived as information characterizing the user. This paper describes an approach for constructing a user signature representing her interests and opinions based on used items and tags. The signature is determined as a fuzzy relation built on two fuzzy sets proposed here: a fuzzy set representing resource attractiveness, and a fuzzy set representing tag popularity. Furthermore, users' signatures are used to determine similarity between users, and potentially give users a method to find new web friends with similar interests and opinions. The paper also describes a process of building different signatures representing a group of users. Signatures of users that are members of the group are aggregated using OWA operator and different linguistic quantifiers to describe the group in a number of ways. A real-world case study illustrating the process of finding similar users and/or groups of users is included.

Semantic Web as an evolution of Web has led to the introduction of new technologies including XML-based formats of representing data on the Web: resource description framework (RDF) and ontology. Similarity assessment of the entities has a fundamental role in processing and analyzing data represented in ontology. In this paper, we propose a technique for determining semantic similarity between pieces of information defined in ontology based on features describing each piece of information. The presented method allows for considering a specific context into the similarity evaluation. The quantitative characterization of similarity at different levels of abstraction in ontology is provided using elements of fuzzy set theory. We show through experiments that the proposed method compares favorably to other measures in terms of human judgment of similarity.

The concept of Semantic Web has introduced an important form of knowledge representation - ontology. As a hierarchical structure of concepts together with their definitions ontology provides means for expressing semantics of data. The ability to build rules with ontology concepts and to perform reasoning increases its attractiveness even further. This paper proposes a framework for expressing fuzzy temporal information using ontology. The framework is built based on ontology suitable for expressing facts and building rules that include fuzzy and temporal terms. This proposed fuzzy temporal ontology can be imported to any domain ontology and used a knowledge base in variety of applications. The paper includes description of build-in predicates needed for constructing fuzzy temporal rules. Simple examples of application of the predicates are presented.

The Internet becomes an enormous source of information containing billions of documents. Users deal with an overwhelming number of alternatives, and continuously make decisions anytime they want to obtain meaningful information. The paper describes an approach for a simple yet effective selection of the most suitable information that fits user's needs. The novelty of the approach is twofold: the concept of lexicographical-like preferences used for a multi-criteria decision-making with elements of fuzziness and OWA operator; a simple fuzzy number based mechanism for estimating user's degrees of acceptance of criterion satisfaction. The lexicographic preferences allow for mimicking user's attitude that some criteria should be satisfied before other criteria are considered. The acceptance of criterion satisfaction levels are defined with a single threshold that represents a boundary value between acceptable and unacceptable values of attributes of alternatives. The paper includes results of a simple case study preformed on a prototype of a web selection system built using the proposed approach.

User's involvement in creating contents of the web is an essential element of Web 2.0 and social software. One example of such involvement is a process called tagging: users annotate digital items, called resources, with labels, called tags. Tags used for annotation represent users' categorization and perception of resources. As the result a network of interconnected resources and tags is created. Such a network is meant to simplify a search process for relevant content. The paper introduced the idea of constructing fuzzy sets based on the network of resources and tags. One of possible ways of building fuzzy representation of tags is proposed and described here. It is shown how those fuzzy sets can be used for search purposes.

The endless amount of information on the web, known as ldquolost-in-hyper-space syndromerdquo, easily overwhelms users. User profiles are used as a means to support extracting relevant information by indicating user interests. In this paper, we propose a new method to develop and maintain a user profile by analyzing user's web access behavior. We propose an ontology-based semantic similarity measure and combine it with an importance measure to identify items that are of highest relevance to user interests. The proposed approach is used in a system for updating a user profile in music domain.

In a significant minority of cases, certain pronouns, especially the pronoun it, can be used without referring to any specific entity. This phenomenon of pleonastic pronoun usage poses serious problems for systems aiming at even a shallow understanding of natural language texts. In this paper, a novel approach is proposed to identify such uses of it: the extrapositional cases are identified using a series of queries against the web, and the cleft cases are identified using a simple set of syntactic rules. The system is evaluated with four sets of news articles containing 679 extrapositional cases as well as 78 cleft constructs. The identification results are comparable to those obtained by human efforts.

The Internet holds huge amount of documents available for users. Effective utilization of this enormous repository means a need for systems supporting users in a process of finding related documents. An ontology defined in the framework of the Semantic Web (Berners, 2001) allows for specification of concepts, their instances, and relationships existing between concepts. A hierarchy of concepts (Yager, 2000) is a graph-like structure providing a means for representing human-like dependencies. The article proposes an approach for utilization of a hierarchy of concepts to perform categorization of web pages in the Semantic Web. A user provides a hierarchy that can only partially “cover” their domain of interest. The hierarchy is treated as a “seed” representing user’s initial knowledge about the domain. Ontologies are treated as supplementary knowledge bases. They are used to instantiate the hierarchy with concrete information, as well as to enhance it with new concepts initially unknown to a user.

Software maintenance is an important phase in the software life cycle. It focuses on keeping the software fully functional and up to date. Maintenance engineers used different approaches and methods to gain understanding of software systems so maintenance tasks can be performed effectively. A lot of efforts have been put into finding a way to measure maintainability of software. It is a common opinion that software maintainability should be described using a set of measurable software attributes. This paper looks at the issue of rule-based description of attributes of software with different levels of maintainability. Varieties of rules are extracted from a data set that represents human evaluation of maintainability of software objects. Rule similarity and rule inclusion measures are used to identify the most diverse sets of rules representing human evaluation criteria. Additionally, the rules representing all evaluators are analyzed using a rule similarity concept in order to learn more about common evaluation criteria

In this paper, we are concerned with the concept of fuzzy logic networks and logic-based data analysis realized within this framework. The networks under discussion are homogeneous architectures comprising of OR/AND neurons originally introduced by Hirota and Pedrycz. Being treated here as generic processing units, OR/AND neurons are neurofuzzy constructs that exhibit well-defined logic characteristics and are endowed with a high level of parametric flexibility and come with significant interpretation abilities. The composite logic nature of the logic neurons becomes instrumental in covering a broad spectrum of logic dependencies whose character spread in-between between those being captured by plain and and or logic descriptors (connectives). From the functional standpoint, the developed network realizes a logic approximation of multidimensional mappings between unit hypercubes, that is transformations from [0, 1]n to [0, 1]m. The way in which the structure of the network has been formed is highly modular and becomes reflective of a general concept of decomposition of logic expressions and Boolean functions (as being commonly encountered in two-valued logic). In essence, given a collection of input variables, selected is their subset and transformed into new composite variable, which in turn is used in the consecutive module of the network. These intermediate synthetic variables are the result of the successive problem (mapping) decomposition. The development of the network is realized through genetic optimization. This helps address important issues of structural optimization (where we are concerned with a selection of a subset of variables and their allocation within the network) and reaching a global minimum when carrying out an extensive parametric optimization (adjustments of the connections of the neurons). The paper offers a comprehensive and user-interactive design procedure including a simple pruning mechanism whose intention is to enhance the interpretability of the network while reducing its size. The experimental studies comprise of three parts. First, we demonstrate the performance of the network on Boolean data (that leads to some useful comparative observations considering a wealth of optimization tools available in two-valued logic and digital systems). Second, we discuss synthetic multivalued data that helps focus on the approximation abilities of the network. Finally, show the generation of logic expressions describing selected data sets coming from the machine learning repository.

Fuzzy cognitive maps (FCMs) form a convenient, simple, and powerful tool for simulation and analysis of dynamic systems. The popularity of FCMs stems from their simplicity and transparency. While being successful in a variety of application domains, FCMs are hindered by necessity of involving domain experts to develop the model. Since human experts are subjective and can handle only relatively simple networks (maps), there is an urgent need to develop methods for automated generation of FCM models. This study proposes a novel evolutionary learning that is able to generate FCM models from input historical data, and without any human intervention. The proposed method is based on genetic algorithms, and is carried out through supervised learning. The paper tests the method through a series of carefully selected experimental studies

Ability to provide convenient access to scientific documents becomes a difficult problem due to large and constantly increasing number of incoming documents and extensive manual work associated with their storage, description and classification. This requires intelligent search and classification capabilities for users to find required information. It is especially true for repositories of scientific medical articles due to their extensive use, large size and number of new documents, and well maintained structure. This research aims to provide an automated method for classification of articles into the structure of medical document repositories, which would support currently performed extensive manual work. The proposed method classifies articles from the largest medical repository, MEDLINE, using state of the art data mining technology. The method is based on a novel associative classification technique which considers recurrent items and most importantly multi-label characteristic of the MEDLINE data. Based on large scale experiments that utilize 350,000 documents several different classification algorithms have been compared including both recurrent and non-recurrent associative classification. The algorithms are capable of assigning each medical document to several classes (multi-label classification) and are characterized by relatively high accuracy. We also investigate different measures of classification quality and point out pros and cons of each. Based on experimental result we show that recurrent item based associative classification demonstrates superior performance and propose three alternative setups that allow the user to obtain different desired classification qualities.

Software maintenance engineers need tools to support their work. To make such tools relevant, they should provide engineers with quantitative input, as well as the knowledge needed to understand factors influencing maintenance activities. This article proposes an approach leading to multitechnique knowledge extraction and development of a comprehensive meta-model prediction system in the area of corrective maintenance. It dwells on elements of evidence theory and a number of fuzzy-based models. The models are developed using an evolutionary-based approach with different objectives applied to different subsets of data. Evidence theory–based Transferable Belief Model and belief function values assigned to generated models are used for reasoning purposes. The study comprises a detailed case for estimating the number of defects in a medical imaging system. © 2005 Wiley Periodicals, Inc. Int J Int Syst 20: 1093–1115, 2005.

Complexity of software is an important aspect of development and maintenance activities. A lot of research is dedicated to defining different software measures that capture what software complexity is. In most cases description of complexity is given to humans in forms of numbers. These quantitative measures reflect human-seen complexity with different levels of success. The paper proposes a process of "translating" human-seen complexity into numbers. The process starts with an experiment that involves human beings and provides data with embedded knowledge about human perception of complexity. Data processing and analysis of data models built based on the data lead to discovery of simple rules, which represent human perception of software complexity.

Although the generation and application of software clones is relatively unexplored, it is believed that this is a fundamental technology that can have many different applications within a software engineering environment. For example, software clones could be used in software fault tolerance. Clearly, for these clones to be usable, their production needs to be automated. An interesting approach to this automatic production or generation problem is the application of evolutionary-based genetic programming (GP). Using the paradigms of best fit, selection, crossover and mutation a number of clones, satisfying specific requirements, can be automatically generated. In general, GP is a flexible and powerful algorithm suitable for solving variety of different problems. The paper presents the results of studies that have been conducted in order to answer questions related to feasibility of using GP for clone generation: what features of GP are important? What works and what does not? How GP can be "tuned" for the problem? The results have been used to draw a set of suggestions and conclusions that indicate possible usability of GP-based approach to automatic generation of clones.

Amount of software engineering data that is accumulated by software companies grows with enormous speed. This data is a source of knowledge about different activities related to software development and maintenance. Many different techniques and tools have been developed and proposed for extracting knowledge and representing it in forms understandable by people. These techniques are based on different principles and they process data differently. This paper illustrates a multi-technique approach to analysis of data. A detailed case study of analyzing software maintenance data is presented. Different models are built, analyzed and evaluated. The first model is a Bayesian network. The second is a set of IF-THEN rules extracted from the data, and the third one is built using a decision tree. The emphasis of the analysis is put on two aspects - how the models support understanding of a process represented by the data, and how good prediction capabilities these models have.

This study is concerned with a general methodology of identification of fuzzy models. Unlike numeric models, fuzzy models operate at a level of information granules - fuzzy sets - and this aspect brings up an important design requirement of transparency of the model. We propose a three-phase development framework by distinguishing between structural and parametric optimization processes. The underlying topology of the model dwells on fuzzy neural networks - architectures governed by fuzzy logic and equipped with parametric flexibility. Two general optimization mechanisms are explored: the structural optimization is realized via genetic programming whereas for the ensuing detailed parametric optimization we proceed with gradient-based learning. The main advantages of this approach are discussed in detail. The study is illustrated with the aid of a numeric example that provides a detailed insight into the performance of the fuzzy models and quantifies crucial design issues.

Effectiveness and clarity of software objects, their adherence to coding standards and programming habits of programmers are important features of overall quality of software systems. This paper proposes an approach towards a quantitative software quality assessment with respect to extensibility, reusability, clarity and efficiency. It exploits techniques of Computational Intelligence (CI) that are treated as a consortium of granular computing, neural networks and evolutionary techniques. In particular, we take advantage of self-organizing maps to gain a better insight into the data, and study genetic decision trees-a novel algorithmic framework to carry out classification of software objects with respect to their quality. Genetic classifiers serve as a "quality filter" for software objects. Using these classifiers, a system manager can predict quality of software objects and identify low quality objects for review and possible revision. The approach is applied to an object-oriented visualization-based software system for biomedical data analysis.

We discuss a problem of rule-based fuzzy modeling of multiple-input single-output nonlinear relationships f: R/sub n//spl rarr/R. The model under investigation is viewed as a collection of conditional statements "if state /spl Omega/, then y=g/sub i/(x,at)", i=1,2,...N with /spl Omega//sub i/ being a fuzzy relation defined in the space of the input variables. In contrast to the commonly encountered identification approach, based exclusively upon discrete experimental data, the one proposed in this study is concerned with the rule-based modeling exploiting the available nonlinear input-output relationship. The main thrust is in the development of a relevant fuzzy partition of the input variables. We introduce and study criteria of separability and variability as the key means guiding a distribution and granularity of the linguistic labels forming the condition part of the local models.