No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

This study investigated the distribution of bibliographic records amongst online bibliographic databases. The topic of Fuzzy Set Theory over the period of 1965 to 1993 was chosen to provide the case study for this investigation. From the DIALOG database host, searches were conducted on 114 databases to determine the number of journal article records relating to the topic of Fuzzy Sets. Both the number of records in each database, as well as the overlap of coverage between the databases were calculated. Six counting techniques were developed to allocate records to databases based on different methods for handling records that were duplicated between databases. When duplicate records are included, the top database accounts for 19% of the records; when duplicates are removed, the top database was found to account for 37% of the records. The distribution of records in databases was found to conform to the Bradford-Zipf hyperbolic distribution. Various other analyses were undertaken including: the duplicate records themselves, the total size of the DIALOG database system over time and the density of Fuzzy Set records in databases over time. A secondary aim of this study was to perform an informetric study on the literature of Fuzzy Set Theory itself. Results obtained include an analysis of the growth of the Fuzzy Set literature, an analysis of the journals covering the topic of Fuzzy Sets, an analysis of the terminology used in describing topics related to Fuzzy Sets. Also, the Ulrich's database was used to provide a subject classification of the journals to analyse the diffusion of the topic of Fuzzy Sets into other disciplines. Apart from the discipline of mathematics, the top disciplines into which Fuzzy Sets have diffused were found to be applied physics, systems and computing. The third aim of the thesis was to refine and develop the methodology used to perform large scale informetric studies using data from a variety of online bibliographic databases. Commercially available software was used wherever possible, but where this was not possible or infeasible, custom written programs were developed to perform various steps in the methodology.

Over the past few years, a considerable number of studies have been made on Support Vector Machines (SVMs) in many domains to improve classification or prediction. However, SVMs request high computational time and memory when the datasets are large. Although incremental learning techniques are viewed as one possible solution developed to reduce the computation complexity of the scalability problem, few studies have considered that some examples close to the decision hyperplane other than support vectors (SVs) might contribute to the learning process. Consequently, we propose three novel algorithms, named Mixed Incremental learning (MIL), Half-Mixed Incremental learning (HMIL), and Partition Incremental learning (PIL), by improving Syed¡¦s incremental learning method based on fuzzy set theory. We expect to achieve better accuracy than other methods. In the experiments, the proposed algorithms are investigated on five standard machine learning benchmark datasets to demonstrate the effectiveness of the method. Experimental results show that HIL have superior classification accuracy than the other incremental or active learning algorithms. Especially, for the datasets that might have high accuracy in other research reports, HMIL and PIL could even improve the performance.

Stastical pattern recognition is the most successful approach to automatic speech and speaker recognition (ASASR). Of all the statistical pattern recognition techniques, the hidden Markov model (HMM) is the most important. The Gaussian mixture model (GMM) and vector quantisation (VQ) are also effective techniques, especially for speaker recognition and in conjunction with HMMs. for speech recognition. However, the performance of these techniques degrades rapidly in the context of insufficient training data and in the presence of noise or distortion. Fuzzy approaches with their adjustable parameters can reduce such degradation. Fuzzy set theory is one of the most, successful approaches in pattern recognition, where, based on the idea of a fuzzy membership function, fuzzy C'-means (FCM) clustering and noise clustering (NC) are the most, important techniques. To establish fuzzy approaches to ASASR, the following basic problems are solved. First, a time-dependent fuzzy membership function is defined for the HMM. Second, a general distance is proposed to obtain a relationship between modelling and clustering techniques. Third, fuzzy entropy (FE) clustering is proposed to relate fuzzy models to statistical models. Finally, fuzzy membership functions are proposed as discriminant functions in decison making. The following models are proposed: 1) the FE-HMM. NC-FE-HMM. FE-GMM. NC-FEGMM. FE-VQ and NC-FE-VQ in the FE approach. 2) the FCM-HMM. NC-FCM-HMM. FCM-GMM and NC-FCM-GMM in the FCM approach, and 3) the hard HMM and GMM as the special models of both FE and FCM approaches. Finally, a fuzzy approach to speaker verification and a further extension using possibility theory are also proposed. The evaluation experiments performed on the TI46, ANDOSL and YOHO corpora showbetter results for all of the proposed techniques in comparison with the non-fuzzy baseline techniques.

In order to store and process natural phenomena in Geographic Information Systems (GIS) it is necessary to model the real world to form computational representation. Since classical set theory is used in conventional GIS software systems to model uncertain real world, the natural variability in the environmental phenomena can not be modeled appropriately. Because, pervasive imprecision of the real world is unavoidably reduced to artificially precise spatial entities when the conventional crisp logic is used for modeling. An alternative approach is the fuzzy set theory, which provides a formal framework to represent and reason with uncertain information. In addition, linguistic variable concept in a fuzzy logic system is useful for communicating concepts and knowledge with human beings. In this thesis, a system to enhance commercial GIS software, namely ArcGIS, with fuzzy set theory is designed and implemented. The proposed system allows users to (a) incorporate human knowledge and experience in the form of linguistically defined variables into GIS-based spatial analyses, (b) handle impreiii cision in the decision-making processes, and (c) approximate complex ill-defined problems in decision-making processes and classification. The operation of the proposed system is presented through case studies, which demonstrate its application for classification and decision-making processes. This thesis shows how fuzzy logic approach may contribute to a better representation and reasoning with imprecise concepts, which are inherent characteristics of geographic data stored and processed in GIS.