http://hdl.handle.net/2139/5462 2013-05-21T20:40:25Z http://hdl.handle.net/2139/10118 Quick fuzzy backpropagation algorithm Nikov, Alexander; Stoeva, Stefka A modification of the fuzzy backpropagation (FBP) algorithm called QuickFBP algorithm is proposed, where the computation of the net function is significantly quicker. It is proved that the FBP algorithm is of exponential time complexity, while the QuickFBP algorithm is of polynomial time complexity. Convergence conditions of the QuickFBP, resp. the FBP algorithm are defined and proved for: (1) single output neural networks in case of training patterns with different targets; and (2) multiple output neural networks in case of training patterns with equivalued target vector. They support the automation of the weights training process (quasi-unsupervised learning) establishing the target value(s) depending on the network's input values. In these cases the simulation results confirm the convergence of both algorithms. An example with a large-sized neural network illustrates the significantly greater training speed of the QuickFBP rather than the FBP algorithm. The adaptation of an interactive web system to users on the basis of the QuickFBP algorithm is presented. Since the QuickFBP algorithm ensures quasi-unsupervised learning, this implies its broad applicability in areas of adaptive and adaptable interactive systems, data mining, etc. applications. 2011-06-14T00:00:00Z http://hdl.handle.net/2139/10117 A methodology for human factors analysis in office automation systems Nikov, Alexander; Matarazzo, Giacinto; Orlando, Antonino A methodology for computer-aided human factors analysis in office automation systems (OAS) design and implementation process has been developed. It incorporates a fuzzy knowledge-based evaluation mechanism which is employed to aggregate data measured in scales of different type. The methodology has a high degree of flexibility which allows it to be adjusted to the individual client situation. A case study in public administration for assessing OAS introduction from users' point of view has been carried out. On the basis of the results recommendations on further implementation have been proposed. The advantages, disadvantages, and further developments are discussed 2011-06-14T00:00:00Z http://hdl.handle.net/2139/10116 NN-AirPol: a neural-networks-based method for air pollution evaluation and control Karaca, Ferhat; Nikov, Alexander; Alagha, Omar A method for air pollution evaluation and control, based on one of the most popular neural networks – the backpropagation algorithm, is proposed. After the backpropagation training, the neural network, based on weather forecasting data, determines the future concentration of critical air pollution indicators. Depending on these concentrations, relevant episode warnings and actions are activated. A case study is carried out to illustrate and validate the method proposed, based on Istanbul air pollution data. Sulphur dioxide and inhalable particulate matter are selected as air pollution indicators (neural network outputs). Relevant episode measures are proposed. Among ten backpropagation algorithms, the BFGS algorithm (Quasi-Newton algorithms) is adopted since it showed the lowest training error. A comparison of NN-AirPol method against regression and perceptron models showed significantly better performance 2011-06-14T00:00:00Z http://hdl.handle.net/2139/10115 A fuzzy backpropagation algorithm Stoeva, Stefka; Nikov, Alexander This paper presents an extension of the standard backpropagation algorithm (SBP). The proposed learning algorithm is based on the fuzzy integral of Sugeno and thus called fuzzy backpropagation (FBP) algorithm. Necessary and sufficient conditions for convergence of FBP algorithm for single-output networks in case of single- and multiple-training patterns are proved. A computer simulation illustrates and confirms the theoretical results. FBP algorithm shows considerably greater convergence rate compared to SBP algorithm. Other advantages of FBP algorithm are that it reaches forward to the target value without oscillations, requires no assumptions about probability distribution and independence of input data. The convergence conditions enable training by automation of weights tuning process (quasi-unsupervised learning) pointing out the interval where the target value belongs to. This supports acquisition of implicit knowledge and ensures wide application, e.g. for creation of adaptable user interfaces, assessment of products, intelligent data analysis, etc. 2011-06-14T00:00:00Z