Hamdi Tolga Kahraman

A User Modeling Approach to Web Based Adaptive Educational Hypermedia Systems

In this paper, a new user modeling approach has been developed to determine the knowledge status of user in Adaptive Educational Hypermedia System (AEHS). The approach is based on forming the domain model and determining relations among elements of that model to decide the knowledge status of user from domain model independently. The proposed method provides flexibility to individual learning or studying steps. In comparison to the literature, the proposed approach enables quick and powerful adaptation for matching instructional needs of users. The difficulties faced in developing such a system are purifying the gathered data, obtaining and evaluating useful data and providing a powerful adaptation effect in a short time.

A novel and powerful hybrid classifier method

Weight-tuning methods and distance metrics have a significant impact on the k-nearest neighbor-based classification. A major challenge is the issue of how to explore the optimal weight values of the features and how to measure distances between the neighbors affecting the classification accuracy of the k-nn. In this paper, a powerful similarity measurement method, which is called the fuzzy distance metric, is explained and extended to measure the distances between the test and training observations. Depending on the fuzzy metric, similarity arrays can be produced more efficiently than the classic and other weighted distance measurements. Finally, the weighting methods are combined with the fuzzy metric-based similarity measurement and the k-nearest neighbor algorithm to increase the classification accuracy of the proposed algorithm. The effectiveness of the proposed approaches is proven by comparing their performances with the performances of the classic and the population-based heuristic methods on the well-known, real-world classification problems obtained from the UCI machine-learning benchmark repository. The experimental results show that the proposed hybrid algorithms significantly explore more optimal weight vectors significantly and provide more accurate classification results than the powerful and well-known instance-based intuitive and heuristic classification algorithms and classic approaches over real datasets. A Novel and Powerful Hybrid Classifier Method has been developed.A novel weight-tuning method is introduced by applying ABC-based heuristic searching approach.A powerful similarity measurement method has been introduced.Experimental results show that the proposed hybrid algorithms significantly improves classification results of the well-known instance-based intuitive and heuristic classification algorithms over real datasets

Application of Adaptive Artificial Neural Network Method to Model the Excitation Currents of Synchronous Motors

In the classic ANN-based approaches, the synchronous motor parameters mostly could be modeled with n-hidden layered networks. It is an important challenge in driver software development is to realize complex mathematical models in real time environments and circuits. This paper presents an Adaptive Artificial Neural Network-based (AANN) method to easily model excitation current of synchronous motors. It has a simple network structure and less processing units (nodes) more than classic ANN. The main purpose of this method are to estimate the excitation current and also to assist designers to model excitation current easily and to develop complex driver software with low degree programming effort while improving the efficiency of classic ANN-based approach. In the adopted approach, the activation functions of nodes in the hidden layers of multilayered feed forward neural network have been determined by using a heuristic method. The experimental results have shown that the proposed method successfully creates single-hidden layered simple networks have less node number than classic ANN-based solutions and achieves the tasks in high estimation accuracies.

Modeling of meteorological parameters and improving the classification accuracy for the wind turbines

In this study novel solutions are presented for challenge subjects such as the weighting of wind energy parameters and classifying of meteorological data that should be consider in the installation of wind turbines. For this purpose, the relationships between the meteorological parameters and wind turbines are explored with an intuitive k-Nearest Neighbor algorithm. Thus, the effects of meteorological data on the power of wind turbines are modeled. In the experimental studies the power class of wind turbines is determined depending on the weighted and unweighted parameters and the performances of various classifiers are compared. The results show that the intuitive classification algorithm determines the power class of wind turbines successfully and produce more correct results than classic k-NN approach.

A Parameter Determination System for Wind Turbines Based on Naive Bayes Classification Algorithm

Among the renewable energy types, wind energy gets popularity in these days. To install a new wind energy turbine, measurement and evaluation of the meteorological data are quite important. In this paper, a novel system is developed for the installation of wind turbines. Firstly, necessary data including the speed and the direction of wind, the solar insolation, the ultraviolet radiation and the rainfall are received from a meteorology station, and then data acquired are converted into useful information using a rule-based inference mechanism. Finally, the useful information obtained are evaluated in a Naive Bayes algorithm. The power and the size of wind turbine are determined automatically using data measured. Thus, some complicated calculations including more than parameters related to fields where the large-sized wind turbines will be installed can easily be accomplished by means of the powerful decision support system developed.

Applying the Meta-heuristic Prediction Algorithm for Modeling Power Density in Wind Power Plant

In this paper, a robust artificial intelligence (AI) algorithm is applied to overcome challenges at power density prediction especially at the installation process of wind power plant. This algorithm also explores relationships between the meteorological parameters and power density. Importance degree of parameters on power density is converted numerical weighting values independently from each other. Thus, the effects of the wind speed, the wind direction, the temperature, the damp, the pressure on power density could be modelled. Besides, experimental study shows that the prediction accuracy and stability of the applied method superior than traditional AI-based techniques.

Optimal scheduling of short-term hydrothermal generation using symbiotic organisms search algorithm

In this study, the Symbiotic Organisms Search (SOS) algorithm is proposed to solve the short-term hydrothermal generation scheduling (STHGS) problem. This problem aims to optimize the power generation strategy produced by hydroelectric and thermal plants by minimizing the total fuel cost function while satisfying some operational constraints. In order to evaluate the effectiveness of the SOS, it has been tested on a system having a hydro plant with four-cascaded reservoir and a thermal plant. Results have been compared other metaheuristic methods. Results obtained from the experiment show that the proposed algorithm produces better results than the other methods and shows a good convergence.

A novel intelligent decision support tool for average wind speed clustering

The utilization ratio of wind energy, which is one of the renewable energy sources, is increased around 25% since last 15 years. However, the parameters such as performance of wind turbines and climate features are not analyzed adequately. At the analysis stage of these parameters, data mining techniques are required to be used. In this study, the agglomerative hierarchical clustering method which is one of the data mining techniques is used to analyze the provinces located in the Central Anatolia Region of Turkey in terms of average wind speed. Nearest neighbor algorithm is used as the clustering algorithm. Euclidean, Manhattan and Minkowski distance metrics are used determine the optimum hierarchical clustering results in this algorithm. The achieved clustering results based on Euclidean distance metric provide the optimum inferences to expert according to other distance metrics.

Exploring the effect of distribution methods on meta-heuristic searching process

In this study, the effect of distributions of solution candidates on the problem space in the meta-heuristic search process and the performance of algorithms has been investigated. For this purpose, solution candidates have been created with random and gauss (normal) distributions. Search performance is measured separately for both types of distribution of algorithms. The performances of the algorithms have been tested on the most popular and widely used benchmark problems. Experimental studies have been conducted on the most recent meta-heuristic search algorithms. It has been seen that the search performance of algorithms varies considerably depending on the method of distribution. In fact, better results were obtained than the distribution methods used in the original versions of the algorithms. Algorithms have revealed their abilities in terms of neighborhoods searching, getting rid of local minimum traps and speeding up searches.

Developing of decision support system for land mine classification by meta-heuristic classifier

In this study, a decision support system has been developed for land mine detection and classification. Data obtained from detector based magnetic anomaly have been used to classify the land mines. With this classification, it is decided that whether obtained data belongs to a land mine or not, and the type of mine. The meta-heuristic k-NN classifier (HKC) has been used in developed decision support system. Consequently, it is seen that decision support system detects the presence of mines and decides the type of mine with 100% success for measurements in a certain range, and the proposed classifying method shows much higher performance than traditional instance-based classification method.