In today&lsquo / s world, knowledge is the most powerful factor for the success of the organizations. One of the most important resources to reach this knowledge is the huge data stored in their databases. In the analysis of this data, DM techniques are essentially used. In this thesis, firstly, a comprehensive literature review on DM techniques for the quality improvement in manufacturing is presented. Then one of these techniques is applied on a case study. In the case study, the customer quality perception data for driver seat quality is analyzed. Decision tree approach is implemented to identify the most influential variables on the satisfaction of customers regarding the comfort of the driver seat. Results obtained are compared to those of logistic regression analysis implemented in another study.

The traveling salesman problem (TSP) is defined as: given a finite number of cities along with the cost of travel between each pair of them, find the cheapest way of visiting all the cities only once and returning to your starting city. Some variants of TSP are proposed to visit cities depending on the profit gained when the visit occurs. In literature, these kind of problems are named TSP with profit. In TSP with profit, there are two conflicting objectives, one to collect profit and the other to decrease traveling cost. In literature, TSP with profit are addressed as single objective, either two objectives are combined linearly or one objective is constrained with a specified bound. In this study, a multiobjective approach is developed by combining &amp / #949 / -constrained method and heuristics from the literature in order to find the efficient frontier for the TSP with profit. The performance of approach is tested on the problems studied in the literature. Also an interactive software is developed based on the multiobjective approach.

In this thesis the radar target tracking problem in Bayesian estimation framework is studied. Traditionally, linear or linearized models, where the uncertainty in the system and measurement models is typically represented by Gaussian densities, are used in this area. Therefore, classical sub-optimal Bayesian methods based on linearized Kalman filters can be used. The sequential Monte Carlo methods, i.e. particle filters, make it possible to utilize the inherent non-linear state relations and non-Gaussian noise models. Given the sufficient computational power, the particle filter can provide better results than Kalman filter based methods in many cases. A survey over relevant radar tracking literature is presented including aspects as estimation and target modeling. In various target tracking related estimation applications, particle filtering algorithms are presented.

Automatic image annotation is the process of assigning keywords to digital images depending on the content information. In one sense, it is a mapping from the visual content information to the semantic context information. In this thesis, we propose a novel approach for automatic image annotation problem, where the annotation is formulated as a multivariate mapping from a set of independent descriptor spaces, representing a whole image, to a set of words, representing class labels. For this purpose, a hierarchical annotation architecture, named as HANOLISTIC (Hierarchical Image Annotation System Using Holistic Approach), is dened with two layers. At the rst layer, called level-0 annotator, each annotator is fed by a set of distinct descriptor, extracted from the whole image. This enables us to represent the image at each annotator by a dierent visual property of a descriptor. Since, we use the whole image, the problematic segmentation process is avoided. Training of each annotator is accomplished by a supervised learning paradigm, where each word is represented by a class label. Note that, this approach is slightly dierent then the classical training approaches, where each data has a unique label. In the proposed system, since each image has one or more annotating words, we assume that an image belongs to more than one class. The output of the level-0 annotators indicate the membership values of the words in the vocabulary, to belong an image. These membership values from each annotator is, then, aggregated at the second layer by using various rules, to obtain meta-layer annotator. The rules, employed in this study, involves summation and/or weighted summation of the output of layer-0 annotators. Finally, a set of words from the vocabulary is selected based on the ranking of the output of meta-layer. The hierarchical annotation system proposed in this thesis outperforms state of the art annotation systems based on segmental and holistic approaches. The proposed system is examined in-depth and compared to the other systems in the literature by means of using several performance criteria.

In this study,we develop an elitist multiobjective evolutionary algorithm for approximating the Pareto-optimal frontiers of multiobjective optimization problems. The algorithm converges the true Pareto-optimal frontier while keeping the solutions in the population well-spread over the frontier. Diversity of the solutions is maintained by the territory de&amp / #64257 / ning property of the algorithm rather than using an explicit diversity preservation mechanism. This leads to substantial computational e&amp / #64259 / ciency. We test the algorithm on commonly used test problems and compare its performance against well-known benchmark algorithms. In addition to approximating the entire Pareto-optimal frontier,we develop a preference incorporation mechanism to guide the search towards the decision maker&amp / #8217 / s regions of interest. Based on this mechanism, we implement two variants of the algorithm. The &amp / #64257 / rst gathers all preference information before the optimization stage to &amp / #64257 / nd approximations of the desired regions. The second one is an interactive algorithm that focuses on the desired region by interacting with the decision maker during the solution process. Based on tests on 2- and 3-objective problems, we observe that both algorithms converge to the preferred regions.

In this study we discuss the Vehicle Routing Problem with multiple use of vehicles (VRPM). In this variant of the routing problem the vehicles may replenish at any time at the depot. We present a detailed review of existing literature and propose two mathematical models to solve the VRPM. For these two models and their several variants we provide computational results based on the test problems taken from the literature. We also discuss a case study in which we are simultaneously dealing with side constraints such as time windows, working hour limits, backhaul customers and a heterogeneous vehicle fleet.

NTRU is a public key cryptosystem, which was first introduced in 1996. It is a ring-based cryptosystem and its security relies on the complexity of a well-known lattice problem, i.e. shortest vector problem (SVP). There is no efficient algorithm known to solve SVP exactly in arbitrary high dimensional lattices. However, approximate solutions to SVP can be found by lattice reduction algorithms. LLL is the first polynomial time algorithm that finds reasonable short vectors of a lattice. The best known attacks on the NTRU cryptosystem are lattice attacks. In these attacks, the lattice constructed by the public key of the system is used to find the private key. The target vector, which includes private key of the system is one of the short vectors of the NTRU lattice. In this thesis, we study NTRU cryptosystem and lattice attacks on NTRU. Also, we applied an attack to a small dimensional NTRU lattice.

Sentiment analysis is the automatic classification of a text, trying to determine the attitude of the writer with respect to a specific topic. The attitude may be either their judgment or evaluation, their feelings or the intended emotional communication. The recent increase in the use of review sites and blogs, has made a great amount of subjective data available. Nowadays, it is nearly impossible to manually process all the relevant data available, and as a consequence, the importance given to the automatic classification of unformatted data, has increased. Up to date, all of the research carried on sentiment analysis was focused on English language. In this thesis, two Turkish datasets tagged with sentiment information is introduced and existing methods for English are applied on these datasets. This thesis also suggests new methods for Turkish sentiment analysis.

Innovations in the public services have become mandatory to provide more efficient and secured services to the citizens. In today&#039 / s fast changing technological environment, the sustained management of innovation is the most vital executive task for the organizations. Identification of the technological innovation process is required in order to manage innovation in the public organizations. This thesis study aims to build a technological innovation model for public organizations in Turkey identifying technological innovation process, stakeholders of the process, sources of innovation, obstacles of innovation and driving forces of innovation. In this research study, strategically important organizations, including all ministries and the pioneer public organizations that perform technological innovation projects are analyzed. In the research study, case study is used as a research strategy and interviews are used as data collection methods. Using collected data / data sets are produced and presented in tables. Data analysis results enable to identify technological innovation process, stakeholders of the process, sources of innovation, barriers of innovation, and driving forces of innovation. Consequently, in accordance with the findings of the study, a new technological innovation model that may pave the way for technological innovation projects and enable successful management of innovation process is constructed. The proposed model lights the way of managers for their innovation projects by means of determining unclear innovation process and identifying the inputs and outputs of the process. Moreover, this study is a guide for managers in public organizations identifying possible obstacles and offering solutions, identifying driving forces to accelerate the innovation process, emphasizing the importance of interaction between the stakeholders.

In this thesis, 2D Local Discriminant Bases (LDB) algorithm is used to 2D search structure to classify remotely sensed data. 2D Linear Discriminant Analysis (LDA) method is converted into an M-ary classifier by combining majority voting principle and linear distance parameters. The feature extraction algorithm extracts the relevant features by removing the irrelevant ones and/or combining the ones which do not represent supplemental information on their own. The algorithm is implemented on a remotely sensed airborne data set from Tippecanoe County, Indiana to evaluate its performance. The spectral and spatial-frequency features are extracted from the multispectral data and used for classifying vegetative species like corn, soybeans, red clover, wheat and oat in the data set.