Article

Fuzzy Classification on Relational Databases

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In practice, information systems are based on very large data collections mostly stored in relational databases. As a result of information overload, it has become increasingly difficult to analyze huge amounts of data and to generate appropriate management decisions. Furthermore, data are often imprecise because they do not accurately represent the world or because they are themselves imperfect. For these reasons, a context model with fuzzy classes is proposed to extend relational database systems. More precisely, fuzzy classes and linguistic variables and terms, together with appropriate membership functions, are added to the database schema. The fuzzy classification query language (fCQL) allows the user to formulate unsharp queries that are then transformed into appropriate SQL statements using the fCQL toolkit so that no migration of the raw data is needed. In addition to the context model with fuzzy classes, fCQL and its implementation are presented here, illustrated by concrete examples. Purchase this chapter to continue reading all 29 pages >

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... The manual definition of membership functions can be applied in expert systems to capture expert knowledge on gradual concepts. This can be applied in information systems for fuzzy classification in databases (Meier, Schindler, & Werro, 2008). When membership functions are explicitly defined a priori, the process of classification corresponds to a mapping from data to membership degrees. ...
... The unsupervised learning of fuzzy classification from data without category labels is called fuzzy cluster analysis (Yang, 1993). In contrast, deductive fuzzy classification means classifying by using membership functions predefined by human experts, e.g., (Meier et al., 2008). Such fuzzy expert systems are based on domain knowledge reflected in membership functions and domain ontologies (Lee & Wang, 2011). ...
... Deductive methods for generating membership function models involve the definition of membership functions by human experts. (Meier et al., 2008) suggest using a membership function editor so that experts can define membership functions graphically. According to Medasani et al. (1998), this is a subjective method for membership function generation based on human perception. ...
Article
Full-text available
Fuzzy classification can be defined as a method of computing the degrees of membership of objects in classes. There are many approaches to fuzzy classification, most of which generate sophisticated multivariate models that classify all of the input space simultaneously. In contrast, methods for membership function generation (MFG) derive simple models for fuzzy classification that map one input variable to one fuzzy class; therefore, by minimizing complexity, these models are very understandable to human experts. The unique contribution of this paper is a method for membership function generation from real data that is based on inductive logic. Most existing MFG methods apply either parameter optimization heuristics or unsupervised learning and clustering for the definition of the membership function. In contrast to heuristic methods, our method can approximate membership functions of any shape. In comparison to clustering, our approach can make use of a target signal to learn a membership function supervised from the association between two variables. Compared to probabilistic methods, which translate frequency information, i.e., normalized histograms, directly into membership degrees, our approach applies inductive reasoning based on conditional relative frequencies, which are called likelihoods. According to the law of likelihood in inductive logic, it is the ratio between the likelihoods of the data that is of interest when evaluating two alternative hypotheses, not the likelihoods themselves. The greatest advantage of our method is its understandability to human users and thereby the potential for visual analytics. However, experimental evaluation did not show reproducible significant effects on the predictive performance of conventional multivariate regression models. Given that there are already many very accurate multivariate models for fuzzy classification, the practical implication is that IFC-Filter can unfold its unique potential mainly for explaining data, specifically, associations between analytical and target variables, to human decision makers. Lessons learned from two case studies with industry partners demonstrate that IFC-Filter can extract interpretable and actionable knowledge from data.
... Fuzzy propositions define fuzzy classes, which allow gradual, fuzzy class boundaries. In data analysis, or "the search for structure in data" (Zimmermann H. J., 1997), fuzzy classification is a method for gradation in data consolidation, as presented by Meier, Schindler, andWerro (2008) andDel Amo, Montero, andCutello (1999). The application of fuzzy classification to marketing analytics (Spais & Veloutsou, 2005) has the advantage of precisiation (sic; Zadeh, 2008) of fuzzy concepts in the context of decision support for direct customer contact, as proposed by Werro (2008). ...
... This thesis, eventually, demonstrates how probabilistic and fuzzy logics can be synthesized to constitute a method of inductive gradual reasoning for classification. 1 1. .1 1 R Re es se ea ar rc ch h Q Qu ue es st ti io on ns s Werro (2008) and Meier et al. (2008) proposed the application of Fuzzy classification to customer relationship management (CRM). A suggestion for further research by Werro (2008), namely the integration of data mining techniques into fuzzy classification software, has inspired the present thesis. ...
... Fuzzy classification is the process of assigning individuals a membership degree to a fuzzy set, based on their degrees of truth of the classification predicate. It has been discussed, for example, by Zimmermann (1997), Del Amo et al. (1999, and Meier et al. (2008). A fuzzy classification is achieved by a membership function, : ⟶ [0,1], that indicates the degree to which an individual is a member of a fuzzy class, , given the corresponding fuzzy propositional function, . ...
... Consequently, inductive fuzzy classification (IFC) is the process of grouping elements into a fuzzy set whose membership function is inferred by induction from data. The induced model for class membership M yields a prediction M(x 1j , …, x nj ) ∈ [0] [1]. ...
... ubtle and smooth distinctions between equivalence classes than crisp classification. This paper presents a methodology for the probabilistic induction of a multivariate fuzzy classification based on likelihood ratios. First, the motivation is a semantically clearly defined automated derivation of fuzzy membership functions for fuzzy classification. [1] and [2] presented the implementation of a classification query language in order to filter databases using linguistic terms. In their research, the question arose how membership functions can be defined using the data. In this paper, a methodology using inductive fuzzy classification is introduced in order to address this issue. Second, ...
... Fuzzy classification ([1], [6], [7]) is the process of grouping elements into a fuzzy set whose membership function is defined by the truth value of a fuzzy constraint predicate. A fuzzy class F = {e | R(e)} is a fuzzy set [8] defined by a fuzzy predicate R(e) where e is an element of a universe of discourse U and R is a fuzzy restriction [9] of U . ...
Conference Paper
Full-text available
A new methodology for the induction of membership functions for fuzzy classification is introduced. The induction step is based on deriving membership functions from normalized likelihood ratios of target class membership. A case study is presented where this fuzzy data analysis process is applied in predictive analytics for an individual marketing campaign in the online channel of a financial service provider by selecting fuzzy target groups in a data warehouse.
... It describes the grade of membership of an individual soil in a defined soil class by matching the environmental variable of the individual soil and the central concept of the soil class. The use of fuzzy membership functions has received particular interest in soil surveys and soil quality assessments (Meier et al., 2008;Dobermann and Oberthür, 1997;Zhu et al., 1997Zhu et al., , 2010Yang et al., 2013;de Gruijter et al., 2011;Beucher et al., 2014;Kaufmann et al., 2009;Nauman et al., 2012). Fuzzy memberships are able to represent the continuous nature of soil spatial variation and are easy to understand for soil experts interested in pedogenesis (Kaufmann et al., 2015). ...
... In general, fuzzy membership classification can be deductive (generated from predefined models) or inductive (constructed from data) (Kaufmann et al., 2015). Deductive membership classification uses membership functions predefined by experts (Meier et al., 2008;Dobermann and Oberthür, 1997;Zhu et al., 1997). Zhu (1997) developed a personal construct-based method and Qi et al. (2006) proposed a prototype-based fuzzy approach to construct fuzzy membership functions for soil-environment knowledge provided by local soil experts. ...
Article
Core Ideas This study develops a method to construct membership functions representing knowledge on soil–environment relationships from partial dependence. Use of representative samples as training samples is recommended when applying the proposed method. Training samples (including representative samples and other samples) with good coverage in the environmental feature space would allow Random Forest to obtain more accurate soil maps than using representative samples. Partial dependence plots generated by Random Forest imply an association between soil and environmental variables. Partial dependence plots generated by Random Forest (RF) imply an association between soil and environmental variables. This study develops a method to construct membership functions representing knowledge of soil–environment relationships from partial dependence. Key parameters were obtained from normalized partial dependence to define class limits and membership gradation. Seven environmental variables were selected on the basis of the variable's importance within RF. Two cases were conducted to test the effectiveness of our method using different training samples. Case 1 used 33 representative locations as training samples and 50 locations as validations. Case 2 randomly split all 83 samples into training and validation subsets at a proportion of 2:1; the splits were repeated seven times. For each case, the generated membership functions were used for mapping soil subgroups in Heshan, China, under the Soil Landscape Inference Model framework; RF was conducted for comparison. The results showed that mapping accuracy based on the membership functions (78%) was much higher than that of RF only (60%) in Case 1. In Case 2, the mapping accuracies using membership functions (an average of 67%, SD = 6.5%) were not always higher than those by RF (an average of 67%, SD = 8.0%). The constructed membership functions were impacted by the training samples. Use of representative training samples is recommended when applying the proposed method. However, training samples (including representative samples and other samples) with good coverage in the environmental feature space would allow RF to obtain more accurate soil maps than using representative samples.
... Many aspects of the present thesis have been published in international conferences , journals and in a handbook. Contributions concerning the fuzzy classication approach, its application to customer relationship management and the implementation aspects have been published in [123, 84, 125, 82, 126, 114, 83, 80]. Since the case study of this thesis has been realized in the e-business eld with online customers, a related topic are online shops and more precisely online shops for SME's. ...
... A logical extension of the mass customization principle is the customization of the products (resp. services) themselves [80]. This means that customers are not choosing a classical product with pre-dened characteristics anymore but a fuzzy product whose specications are automatically adapted to the customer's taste and/or requirements. ...
Book
Abstract The following,paper,presents a case study conducted,for a non profit or- ganisation (BSU). BSU is a student organisation from Fribourg, Switzerland that regularly organises stock market,simulations,for students. The purpose of this case study was to analyse participants using a hierarchal fuzzy clas- sification. After a brief overview of BSU, di erent methods for analysing online customers,are analysed. Fuzzy set are presented,next and finally a hierarchical fuzzy classification of online participants is presented.
... Based on the approximate reasoning, Subsection presents the fuzzy control theory in comparison to the modern control theory. Examples of fuzzy diagnosis and fuzzy data analysis areas are fuzzy expert systems depicted in Subsection and the fuzzy classification approach presented [9] Last but not least, Subsection presents different approaches which enable the representation and the storage of the imprecision, i.e. fuzzy databases systems. Nowadays a large number of real-world applications take advantages of the approximate reasoning. ...
Article
Full-text available
This thesis has been realized following a design science approach, it therefore aims at first creating innovative concepts which improve the actual human and organizational capabilities, secondly, at evaluating these concepts by providing concrete instantiations. According to this research paradigm, the objectives of this thesis are the following: The first objective of this thesis is to extend the querying ability of the fuzzy classification approach proposed by Schindler. By adding new clauses to the fuzzy Classification Query Language, the user should be given more powerful means for selecting elements within a fuzzy classification. The second objective of this thesis is convert classical value to fuzzy value which base on fuzzy membership function such as S shape membership function, Pi shape membership function and Z shape membership junction. The third objective is, considering the application domain specificities, to extend the original fuzzy queries approach by new concepts which provide additional capabilities to the system and proved that the proposed intelligent fuzzy query is faster than the conventional query and it provides the user the flexibility to query the database using natural language. The fourth and last objective is to also make a comparison between traditional database and fuzzy database by computing the time cost of classical query over classical database, fuzzy query over classical database and fuzzy query over fuzzy database.
... Fuzzy association analysis computes association rules between fuzzy restrictions on variables. Fuzzy classification partitions sharp data into fuzzy sets according to a classification predicate (Meier, Schindler, & Werro, 2008). If this predicate is inferred by induction, the process is called inductive fuzzy classification (IFC) by Kaufmann & Meier (2009). ...
Chapter
Full-text available
Scoring models yield continuous predictions instead of sharp classifications. Scoring customers for profitability, loyalty, or product affinity corresponds to an inductive fuzzy classification: The model represents a continuous membership function mapping the set of customers into the fuzzy set of interesting customers – the fuzzy target group. This chapter presents a method for membership function induction based on normalized likelihood ratios. Applications of this method are proposed for selection, visualization, and prediction in the field of analytics in general, and for customer profiling, target group definition and customer scoring specifically for analytic customer relationship management. A real world case study is described. Furthermore, an implementation of the proposed method, developed at the research center for fuzzy marketing methods (FMsquare1), is presented.
... One of the fundamentals of marketing science is that customer behavior cannot be claimed to be well understood until it can be detailed into quantitative terms [21]. Fuzzy logic set effectively handles vague, inexact, stochastic input variables, and treats the dynamic nature of such variables. ...
... A linguistic variable (LV) is defined in Def. 2 [138]. Definition 2: A linguistic variable (LV) is characterized by a quintuple (x, T, X, L, M ), where x is the variable, T is the set of linguistic terms of x, X is an universe of discourse, L is a syntactic rule for generating the terms, and M is a semantic rule to associate each term with its meaning. ...
Article
Full-text available
Aggregation function is an important component in an information aggregation or information fusion system. Interrelationships usually exist between the input arguments (e.g. the criteria in the multi-criteria decision making) of an aggregation function. In this paper, we make a comprehensive survey on the aggregation operators (AOs) that consider the argument interrelationships in crisp and fuzzy settings. In particular, we discuss the mechanisms of modelling the argument interrelationships of the Choquet integral (CI), the power average (PA), the Bonferroni mean (BM), the Heronian mean (HM) and the Maclaurin symmetric mean (MSM) operators, and introduce their extended (e.g. generalized or weighted) forms and their applications in different fuzzy sets. In addition, we compare these five types of operators and summarize their advantages and disadvantages. Furthermore, we discuss the applications of these operators. Finally, we identify some future research directions in the AOs considering the argument interrelationships. The reviewed papers are mainly about the development of the CI, the PA, the BM, the HM and the MSM in (fuzzy) MCDMs, most of which fall in the period of 2009-2018.
... This explains why we must build more complicated schemes than simple partitions or hierarchies. In many situations, fuzzy models (in the sense of Zadeh 1965) extended now to big relational data bases (Meier et al. 2008), or rough sets (in the sense of Pawlak 1982) are necessary, because an object may belong more or less to some class. And a cloudy organization is sometimes better than none. ...
Article
One of the main topics of scientific research, classification is the operation consisting of distributing objects in classes or groups which are, in general, less numerous than them. From Antiquity to the Classical Age, it has a long history where philosophers (Aristotle), and natural scientists (Linnaeus), took a great part. But from the nineteenth century (with the growth of chemistry and information science) and the twentieth century (with the arrival of mathematical models and computer science), mathematics (especially theory of orders and theory of graphs or hypergraphs) allows us to compute all the possible partitions, chains of partitions, covers, hypergraphs or systems of classes we can construct on a domain. In spite of these advances, most of classifications are still based on the evaluation of ressemblances between objects that constitute the empirical data. However, all these classifications remain, for technical and epistemological reasons we detail below, very unstable ones. We lack a real algebra of classifications, which could explain their properties and the relations existing between them. Though the aim of a general theory of classifications is surely a wishful thought, some recent conjecture gives the hope that the existence of a metaclassification (or classification of all classification schemes) is possible. © 2018 International Society for Knowledge Organization. All rights reserved.
... It takes the subjectivity, imprecision, uncertainty and vagueness of human thinking and language into account, and expresses it with mathematical functions (Zadeh 1965). A fuzzy set can be defined formally as follows (Zimmermann 1991, Werro 2008, Meier et al. 2008: if X is a set, then a fuzzy set A in X (A ⊂ X) is defined as ...
Chapter
Full-text available
In the Internet economy and information society, it has become an essential task of electronic business to analyze, to monitor, and to optimize websites and Web offers. Therefore, this chapter addresses the issues of Web analytics, which is defined as the measurement, collection, analysis, and reporting of Internet data for the purposes of understanding and optimizing website usage. After a short introduction, the second section defines Web analytics, describes benefits and problems of Web analytics, as well as different software architectures and products. Third, a controlling loop is proposed for Web content and Web user controlling in order to analyze Key Performance Indicators (KPIs) and to take website-and e-business-related actions. Fourth, different Web metrics and KPIs of information, transaction and communication are defined. Fifth, a fuzzy Web analytics approach is proposed, which makes it possible to classify Web metrics precisely into more than one class at the same time. Considering real Web data of the Web metrics page views and bounce rate, it is shown that fuzzy classification allows exact and flexible segmentation of Web pages or other objects and gradual rankings within fuzzy sets. In addition, the fuzzy logic approach enables Computing with Words (CWW), i.e. the perception-based, linguistic consideration of Web data and Web metrics instead of measurement-based, numerical ones. Web usage mining with inductive fuzzy classification and Web Analytics with Words (WAW) allows intuitive, human-oriented analyses, description, and reporting of Web metrics values in natural language.
... 50 It has been discussed by numerous authors. [51][52][53] In this work, subtractive clustering technique was used to generate a fuzzy inference system (FIS). 54 The clustering technique examines the dataset and establishes both number of clusters and cluster centers. ...
... Fuzzy Sugeno classifier (FSC) classification is based on the process of grouping elements into a FSC set [38] whose membership function was defined by the truth value of a FSC propositional function [39]. This technique has been discussed by numerous authors [40][41][42]. In this work, we have used the subtractive clustering technique to assemble a fuzzy inference system (FIS) [43]. ...
... First, the motivation is a semantically clearly defined automated derivation of fuzzy membership functions for fuzzy classification. [1] and [2] presented the implementation of a classification query language in order to filter databases using linguistic terms. In their research, the question arose how membership functions can be defined using the data. ...
Conference Paper
Full-text available
Inductive fuzzy classification (IFC) is the process of grouping elements into fuzzy sets whose membership functions are induced from data. These methods can be applied to fuzzy data analysis for knowledge discovery, reporting and prediction. In this paper, two new IFC methods using percentile ranks and likelihood ratios and their application to fuzzy data analysis in general and fuzzy marketing analytics in specific are investigated. The benefit of IFC to customer profiling, product scoring, campaign measurement and web usage mining is discussed.
... The core power of fuzzy logic is the fuzzy set theory, first proposed by Zadeh as an extension of the traditional set theory [22]. Meier et al. describe fuzzy logic as a suitable instrument for roughly modeling the kind of uncertainty related to vagueness [12]. Fuzzy logic is an addition to conservative logic and handles the concept of partial truth along with true and false, which is used for qualitative rather than quantitative judgment. ...
... Fuzzy association analysis computes association rules between fuzzy restrictions on variables. Fuzzy classification partitions sharp data into fuzzy sets according to a classification predicate (Meier, Schindler, & Werro, 2008). If this predicate is inferred by induction, the process is called inductive fuzzy classification (IFC) by Kaufmann & Meier (2009). ...
Chapter
Full-text available
Scoring models yield continuous predictions instead of sharp classifications. Scoring customers for profitability, loyalty, or product affinity corresponds to an inductive fuzzy classification: The model represents a continuous membership function mapping the set of customers into the fuzzy set of interesting customers – the fuzzy target group. This chapter presents a method for membership function induction based on normalized likelihood ratios. Applications of this method are proposed for selection, visualization, and prediction in the field of analytics in general, and for customer profiling, target group definition and customer scoring specifically for analytic customer relationship management. A real world case study is described. Furthermore, an implementation of the proposed method, developed at the research center for fuzzy marketing methods (FMsquare1), is presented.
... A FDWH can be queried on a linguistic level. For example, fCQL (Meier et al. [10]) allows marketers to classify single customers or customer groups by classification predicates such as 'loyalty is high and turnover is large'. ...
Conference Paper
Full-text available
In classical data warehouses (DWH), classification of values takes place in a sharp manner, because of this true values cannot be measured and smooth transition between classes does not occur. In this paper, a fuzzy data warehouse (FDWH) modeling approach, which allows integration of fuzzy concepts without affecting the core of a DWH is presented. This is accomplished through the addition of a meta-table structure, which enables integration of fuzzy concepts on dimensions and facts, while preserving the time-invariability of the DWH and allowing analysis of data both sharp and fuzzy. A comparison to existing approaches for integrating fuzzy concepts in DWH is presented. Guidelines for modeling the fuzzy meta-tables and a meta-model for the FDWH are also outlined in this paper. The use of the proposed approach is demonstrated by a retail company example. Finally, a comparison of fuzzy and classical data warehousing approaches is presented.
... A fuzzy set can be defined formally as follows (1; Zimmermann 1992, Meier et al. 2008: if X is a set, then a fuzzy set A in X (A X) is defined as ...
... This paper discusses an approach to find more appropriate information by searching weblogs using fuzzy logic, here referred to as fuzzy weblog extraction . In [1], Meier, Schindler and Werro describe fuzzy logic as an appropriate instrument for rough modelling the kind of uncertainty related with vagueness. The core power of fuzzy logic is the fuzzy set theory, first proposed by Zadeh in [2] as an extension of the traditional set theory. ...
Article
Full-text available
This paper presents fuzzy clustering algorithms to establish a grassroots ontology - a machine-generated weak ontology - based on folksonomies. Furthermore, it describes a search engine for vaguely associated terms and aggregates them into several meaningful cluster categories, based on the introduced weak grassroots ontology. A potential application of this ontology, weblog extraction, is illustrated using a simple example. Added value and possible future studies are discussed in the conclusion.
Chapter
This chapter is dedicated to post-relational database aspects, which are database features that go beyond the traditional relational model. These databases include federal databases, temporal databases, and multi-dimensional databases. Also, the chapter introduces data warehouses and data lakes, which combine federal, temporal, and multi-dimensional properties. Moreover, object-oriented extensions in object-relational databases are discussed. Additionally, knowledge-based and fuzzy-based databases are mentioned.
Chapter
Text Analysis with Python: A Research-Oriented Guide is a quick and comprehensive reference on text mining using python code. The main objective of the book is to equip the reader with the knowledge to apply various machine learning and deep learning techniques to text data. The book is organized into eight chapters which present the topic in a structured and progressive way. Key Features · Introduces the reader to Python programming and data processing · Introduces the reader to the preliminaries of natural language processing (NLP) · Covers data analysis and visualization using predefined python libraries and datasets · Teaches how to write text mining programs in Python · Includes text classification and clustering techniques · Informs the reader about different types of neural networks for text analysis · Includes advanced analytical techniques such as fuzzy logic and deep learning techniques · Explains concepts in a simplified and structured way that is ideal for learners · Includes References for further reading Text Analysis with Python: A Research-Oriented Guide is an ideal guide for students in data science and computer science courses, and for researchers and analysts who want to work on artificial intelligence projects that require the application of text mining and NLP techniques.
Chapter
Nowadays, big data is available in every field due to the advent of computers and electronic devices and the advancement of technology. However, analysis of this data requires new technology as the earlier designed traditional tools and techniques are not sufficient. There is an urgent need for innovative methods and technologies to resolve issues and challenges. Soft computing approaches have proved successful in handling voluminous data and generating solutions for them. This chapter focuses on basic concepts of big data along with the fundamental of various soft computing approaches that give a basic understanding of three major soft computing paradigms to students. It further gives a combination of these approaches namely hybrid soft computing approaches. Moreover, it also poses different applications dealing with big data where soft computing approaches are being successfully used. Further, it comes out with research challenges faced by the community of researchers.
Chapter
Marketing deals with identifying and meeting the needs of customers. It is therefore both an art and a science. To bridge the gap between art and science, soft computing, or computing with words, could be an option. This chapter introduces fundamental concepts such as fuzzy sets, fuzzy logic, and computing with linguistic variables and terms. This set of fuzzy methods can be applied in marketing and customer relationship management. In the conclusion, future research directions are given for applying fuzzy logic to marketing and customer relationship management.
Chapter
Chapter 6 is dedicated to postrelational databases such as federal, temporal, multidimensional databases as well as data warehousing, object-relational, knowledge-based and fuzzy-based databases.
Chapter
This chapter presents a case study in performance measurement using a movie rental company. The case study aims to point out the advantages of a fuzzy data warehouse when classifying elements in a data warehouse. First, the movie rental company and its initial, classical data warehouse are presented in Sect. 4.1. In Sect. 4.2, fuzzy concepts are applied to the data warehouse in order to build a fuzzy data warehouse.
Chapter
As shown in Chap. 4, the application of fuzzy concepts for analysis can provide better results than classical analysis. The interpretation of measures in non-numeric, meaningful linguistic terms provide a more accurate interpretation of measures for business users. In [FZ09], a fuzzy data warehouse was proposed for a web analytics system.
Chapter
Data warehouse was first discussed by Devlin and Murphy in 1988 [DM88]. They described a read-only database for integration of historical operation data and propose tools for user interaction with this database for decision support and analysis. However, Inmon’s definition has received the most attention over the years. According to Inmon [Inm05], “a data warehouse is a subject oriented, non volatile, integrated and time variant collection of data in favor of decision making”.
Chapter
In this chapter, a fuzzy data warehouse concept based on meta table structure is presented. To start with, existing approaches are presented and analyzed in Sect. 3.1. The different fuzzy data warehouse approaches are then compared and evaluated in Sect. 3.1.4. In order to address the problems identified in Sect. 3.1.4, Sect. 3.2 presents a new fuzzy data warehouse concept based on a meta table structure.
Chapter
Chapter 4 discussed the application of a fuzzy data warehouse for a movie rental company, and demonstrated how a fuzzy concept can be integrated in data analysis. The corresponding SQL statements and result set were shown. However, for end users, the application only provides direct access to the database system of the data warehouse. Therefore, the user has to know the structure of the meta tables and the data warehouse.
Chapter
Solving problems in place, on an individual, organizational or even global level comes down to methodical access to knowledge. Networks are so vital in order to enable developing new and innovative solutions that cultivate and disseminate collective knowledge. Networks of people and organizations must necessarily be aligned with networks of knowledge. While knowledge is generally highly cross-linked, at times, this cross-linking of Social Web data is still hard to see. As a consequence, today’s social media elements can prove impractical for the expansion of new and innovative solutions. So the lack of this cross-linking can hinder elementary information management and problem-solving potentials, such as finding, creating and deploying the right knowledge at the right time. Accordingly, a semantic extension of the Social Web is highly aimed at since among social media elements sometimes only little knowledge is exchanged. Unfortunately, precisely this sparse knowledge exchange can lead to redundant knowledge bases, which in turn affects the problem of information overload.
Article
Full-text available
Curriculum development, renewal and innovation are critical to successful engineering programs. In this paper, the authors describe the development of a new mechatronic engineering program at a new college of Higher Education. Small class sizes and high staff-student ratio are among the distinguishing characteristics of the college. The development of a mechatronic degree was taken to ensure that it meets the requirements of the Ministry of Defence Engineering Services. It has been both innovative and strategically designed in terms of balancing between the technical competencies required and the academic rigor of the 21st century engineering programs meeting international standards. The resulting program curriculum is thus both flexible and relevant. The flexibility of design of the Systems Engineering programs allows the pursuit of a variety of pathways. In addition, the curriculum is relevant as it is structured to allow for emphasis areas, aligned with stakeholders’ requirements, ensuring fit for purpose knowledge and minimizing the period of the on job training. The paper discusses the resulting implementation of the course and highlights the analysis and mapping against the expected profile of the graduates. Keywords: mechatronics, skills-set, multidisciplinary engineering, training needs analysis, learning outcomes, engineering education, fuzzy classification
Chapter
Postrelationale Datenbanksysteme sind Erweiterungen klassischer relationaler Softwareprodukte. Mit Replikaten oder Fragmenten lassen sich föderierte Datenbanksysteme entwickeln. Eine besondere Herausforderung bildet die Gültigkeitszeit, die bei temporalen Datenbanken Anwendung findet. Ein Data Warehouse besteht im Kern aus einem mehrdimensionalen Datenwürfel, wobei die gespeicherten Kennzahlen entlang unterschiedlicher Auswertungsdimensionen analysiert werden können. Neben objektrelationalen und wissensbasierten Datenbanken versucht man mit der Hilfe von unscharfen Methoden (Fuzzy Logic) und linguistischen Ansätzen, unpräzise oder vage Sachverhalte auszuwerten. Dieser Lösungsansatz gewinnt bei Big Data an Bedeutung.
Chapter
Nowadays the Web is omnipresent, reaching into almost everyone’s life. More and more Web users do not switch off their devices all the time, continuously receiving and sending messages, frequently looking for information, now and then evaluating this information, and so on. The means to reach the Web do thereby not stop at personal computers, but increasingly also include mobile devices. More and more users are sharing information online, are working collaboratively on a topic, as well as maintaining their relationship in the Web (Alby, 2008). All of this is so pervasive that it feels absolutely natural. Consequently it is not surprising that topics related to the Social Web are experiencing a surge of interest, both from the scientific community as well as the industry. However, apart from this and maybe also apart from the public perception, a complementary technological revolution takes place—the rising adaption of Semantic Web technologies. The Semantic Web is a vision that the present Web will eventually include the notion of meaning and become a metadata-rich Web where presently human-readable content will contain computer-understandable semantics (Berners-Lee, Hendler, & Lassila, 2001).
Chapter
Most of the conventional tools for formal modeling, reasoning, and computing are hard, deterministic, and precise. Thereby hard implies unambiguity that is, yes-or-no rather than more-or-less. In traditional bivalent logic, for example, a statement can be true or false—and nothing in-between. Precision assumes that parameter of a model typifies precisely the features of a real system that has been modeled. Usually, precision also implies that a model is doubtless, that is, that it covers no ambiguities (Zimmermann, 2001).
Chapter
This chapter examines the foundations of IFC by analyzing the concepts of deduction, fuzziness, and induction. The first subsection explains the classical concepts of sharp and deductive logic and classification; in this section, it is presupposed that all terms are clearly defined. The second section explains what happens when those definitions have fuzzy boundaries and provides the tools, fuzzy logic and fuzzy classification, to reason about this. However, there are many terms that do not only lack a sharp boundary of term definition but also lack a priori definitions. Therefore, the third subsection discusses how such definitions can be inferred through inductive logic and how such inferred propositional functions define inductive fuzzy classes. Finally, this chapter proposes a method to derive precise definitions of vague concepts—membership functions—from data. It develops a methodology for membership function induction using normalized likelihood comparisons, which can be applied to fuzzy classification of individuals.
Chapter
Analytics is “the method of logical data analysis” (merriam-webster.com, 2012a). According to Zimmermann (1997), data analysis is the “search for structure in data”. The more data is available, the more complex it becomes to find relevant information. Consequently, organizations and individuals analyze their data in order to gain useful insights. Business analytics is defined as “a broad category of applications and techniques for gathering, storing, analyzing and providing access to data to help enterprise users make better business and strategic decisions” (Turban, Aronson, Liang, & Sharda, 2007, p. 256).
Chapter
How many grains does it take to constitute a heap? This question is known as the sorites paradox (Hyde, 2008). It exemplifies that our semantic universe is essentially vague, and with any luck, this vagueness is ordinal and gradual. This applies to all kinds of statements. Especially in science, different propositions or hypotheses can only be compared to each other with regard to their relative accuracy or predictive power. Fuzziness is a term that describes vagueness in the form of boundary imprecision.
Article
In order to apply naïve Bayes classifiers in more complex real-world situations, we design a newly hybrid classifier that is an ensemble of classifiers. The ensemble classifier is constructed by incorporating hard and fuzzy classification techniques such as fuzzy c-means clustering and naïve Bayes classification. The fuzzy c-means clustering works the preprocessing step, which generates a fuzzy partition based on a given propositional function to augment the above ensemble classifier. This strategy would work better than a conventional hard classifier without fuzzy classification. Our experimental results show the newly hybrid classifier has improved the accuracy and stability of classification, and its classification performance is much closer to that of the tree augment naïve Bayes classifier. In a word, the mathematical work performed in the newly hybrid classifier is not only theoretically admirable, but it also works in many practical applications.
Article
Marketing deals with identifying and meeting the needs of customers. It is therefore both an art and a science. To bridge the gap between art and science, soft computing, or computing with words, could be an option. This chapter introduces fundamental concepts such as fuzzy sets, fuzzy logic, and computing with linguistic variables and terms. This set of fuzzy methods can be applied in marketing and customer relationship management. In the conclusion, future research directions are given for applying fuzzy logic to marketing and customer relationship management.
Article
Full-text available
In this paper, we present a novel hybrid classification model with fuzzy clustering and design a newly combinatorial classifier for error-data in joining processes with diverse-granular computing, which is an ensemble of a naïve Bayes classifier with fuzzy c-means clustering. And we apply it to improve classification performance of traditional hard classifiers in more complex real-world situations. The fuzzy c-means clustering is applied to a fuzzy partition based on a given propositional function to augment the combinatorial classifier. This strategy would work better than a conventional hard classifier without fuzzy clustering. Proper scale granularity of objects contributes to higher classification performance of the combinatorial classifier. Our experimental results show the newly combinatorial classifier has improved the accuracy and stability of classification.
Article
It is difficult to detect and treat alcoholism, because statistics show that statements from patients about their drinking habits are unreliable and diagnosable symptoms appear only in advanced stages of the disease. To address this problem, we propose an automatic system that characterizes alcohol related abnormalities in Electroncephalography (EEG) signals. This system enables clinicians, patients and all other people involved to manage the condition better. Furthermore, it provides deeper insights into the phenomena and thereby it reveals important clinical information about alcohol related changes in EEG signals. For this work, we adopt the widely held, and evidence supported, belief that EEG recordings are fundamentally nonlinear. As direct consequence, the nonlinear feature of Higher Order Spectra (HOS) cumulants was used to extract information about alcohol related changes from the EEG signals. The decision whether or not a particular EEG signal shows alcohol related changes, was established with six different classification algorithms: Decision Tree (DT), Fuzzy Sugeno Classifier (FSC), K-Nearest Neighbor (KNN), Gaussian Mixture Model (GMM), Naive Bayes Classifier (NBC) and Probabilistic Neural Network (PNN). To establish the functionality, we tested the proposed diagnosis support system with 300 EEG data sets. The individual classification algorithms achieved different accuracy values, they ranged from 77% (NBC) to 92.4% (FSC). The (FSC) classification result supports our thesis that HOS based cumulants features can be used to discriminate alcohol and normal EEG signals. The fact that there was a wide range of classification accuracies supports our decision to test four different classification algorithms.
Article
In this study, we investigate the use of morphometric parameters and fuzzy membership functions to perform landform classification for different case areas of Zagros Mountains from digital elevation models (DEMs). First, multiscale DEMs with scales of 5 to 45 cells are generated using the lifting scheme. The maximum curvature for the scale of five cells has the lowest standard deviation, and hence, is determined to be the characteristic scale. Data layers are produced from the DEM of this scale for slope, minimum and maximum curvatures, and plan and profile curvatures. The fuzzy membership rules for these data layers are used to determine the landform classes. Comparison of the results of landform classification using the fuzzy classification method and topographic position index (TPI) with the geology map of the study area show that the fuzzy classification method provides higher accuracy (81 %) as compared to TPI (42 %). This is because for the fuzzy classification method, sloping areas are separated into sloping and non-sloping areas, and the membership functions are defined to prevent landforms belonging to the sloping areas from being classified in the non-sloping areas and vice versa.
Article
This paper describes a computer-based identification system of normal and alcoholic Electroencephalography (EEG) signals. The identification system was constructed from feature extraction and classification algorithms. The feature extraction was based on wavelet packet decomposition (WPD) and energy measures. Feature fitness was established through the statistical t-test method. The extracted features were used as training and test data for a competitive 10-fold cross-validated analysis of six classification algorithms. This analysis showed that, with an accuracy of 95.8%, the k-nearest neighbor (k-NN) algorithm outperforms naive Bayes classification (NBC), fuzzy Sugeno classifier (FSC), probabilistic neural network (PNN), Gaussian mixture model (GMM), and decision tree (DT). The 10-fold stratified cross-validation instilled reliability in the result, therefore we are confident when we state that EEG signals can be used to automate both diagnosis and treatment monitoring of alcoholic patients. Such an automatization can lead to cost reduction by relieving medical experts from routine and administrative tasks.
Article
Die relationale Datenbanktechnologie hat sich in den letzten Jahren breit im Markt durchgesetzt. Ein Ende dieser erfolgreichen Entwicklung ist noch nicht abzusehen. Trotzdem stellt sich die Frage, wohin die Reise führt. Da ist die Rede von verteilten Datenbanksystemen, von temporalen, von deduktiven, von semantischen, von objektorientierten Systemen, von unscharfen, von versionenbehafteten etc. Was verbirgt sich hinter all diesen schillernden Adjektiven? Das vorliegende Kapitel erläutert einige dieser Begriffe und zeigt künftige Methoden und Entwicklungstendenzen auf, wobei die Auswahl subjektiv bleiben muss.
Article
This study is aimed to utilize fuzzy classification for different object detections that is urban, infrastructure and coastal water in RADARSAT-1 SAR S2 mode data. Prior to fuzzy classification, Lee algorithm with kernel window sizes of 7 × 7 pixels and lines is implemented to S2 mode data. Indeed, speckle reduction is performed using Lee algorithm. The results show that Lee algorithm is able to provide excellent information about linear infrastructure and urban features in SAR data. Further, fuzzy classification can discriminate between urban zone and coastal waters. In conclusion, the integration between Lee algorithm and fuzzy classification can be used for different object recognitions in S2 mode data.
Technical Report
Full-text available
Entscheidungsfragen lassen sich bei anspruchsvollen Managementaufgaben nicht immer scharf mit ja oder nein beantworten. Vielmehr geht es um ein Abwägen unterschiedlicher Einflussfaktoren und die Antwort für eine Problemlösung lautet oft ‚ja unter Vorbehalt’ oder ‚sowohl als auch’. Die Antwort ist unscharf und kann Werte zwischen 0 und 1 annehmen. Die unscharfe Logik entspricht der menschlichen Wahrnehmung. Sie vermag neben quantitativen Größen qualitative Einschätzungen mit einzubeziehen. Um Entscheidungsfindung bei vagem Sachverhalt in Informationssystemen zu ermöglichen, müssen Managementmethoden mit unscharfen Konzepten erweitert werden. Der Beitrag führt in die unscharfe Logik ein und zeigt deren Potenzial anhand der unscharfen Scoringmethode fRFM (fuzzy Recency-, Frequency- und Monetary-Werte) auf, die beim schweizerischen Detailhändler coop@home testweise angewendet wurde.
Conference Paper
This paper uses a fuzzy data warehouse approach to support the fuzzy analysis of the customer performance measurement. A data warehouse concept supporting fuzzy dimensions and fuzzy facts is described. The potential of the fuzzy data warehouse approach is illustrated using a concrete example of customer performance measurements of a hearing instrument manufacturer. The example discusses the creation of fuzzy multidimensional classification spaces with using the dicing operation and demonstrate the added value of fuzzy slices, dices and aggregations compared to crisp ones. Added value and potential future studies are also discussed in the conclusion.
ResearchGate has not been able to resolve any references for this publication.