Browsing by Subject "Neural networks (Computer science)"

Now showing 1 - 4 of 4

Open Access
Approximating small open economy models with neural network trained by genetic algorithm
(2008) Coşkun, Yeşim
This thesis work presents a direct numerical solution methodology to approximate the small open economy models with debt elastic interest rate premium and with convex portfolio adjustment cost, both studied by Stephanie SchmittGroh´e and Martin Uribe(2003). This recent method is compared with the firstorder approximation to the policy function from the aspect of second moments of endogenous variables and their impulse responses. The proposed methodology, namely genetic algorithm-neural network (GA-NN), parameterizes the policy function with a feed-forward neural network that is trained by a genetic algorithm. Thus, unlike the first-order approximation, GA-NN does not require the continuity and the existence of derivatives of objective and policy functions. Importantly, since genetic algorithm is an evolutionary algorithm that enables global search over the feasible set, it provides a robust result in any solution space. Also GA-NN method gives not only the moments of the model but also the optimal path.
Open Access
Artificial neural networks modeling and simulation of the in-vitro nanoparticles - cell interactions
(2012) Cenk, Neslihan
In this research a prediction model for cellular uptake efficiency of nanoparticles (NPs), which is the rate of NPs adhered to the cell surface or entered into the cell, is investigated via Artificial Neural Network (ANN) method. Prediction of cellular uptake rate of NPs is an important study considering the technical limitations of volatile environment of organism and the time limitation of conducting numerous experiments for thousands of possible variations of different variables that have an impact on NP uptake rate. Moreover, this study constitutes a basis for the targeted drug delivery and cell-level detection, treatment and diagnoses of existing pathologies through simulating experimental procedure of NP-Cell interactions. Accordingly, this study will accelerate nano-medicine researches. The research focuses on constructing a proper ANN model based on multilayered feed-forward back-propagation algorithm for prediction of cellular uptake efficiency which depends on NP type, NP size, NP surface charge, concentration and time. NP types for in-vitro NP-healthy cell interaction analysis are polymethyl methacrylate (PMMA), silica and polylactic acid (PLA) all of whose shapes are spheres. The proposed ANN model has been developed on MATLAB Programming Language by optimizing number of hidden layers, node numbers and training functions. The data sets for training and testing of the network are provided through in-vitro NP-cell interaction experiments conducted by a Nano-Medicine Research Center in Turkey. The dispersion characteristics and cell interactions of the different nanoparticles in organisms are explored through constructing and implementing an optimal prediction model using ANNs. Simulating the possible interactions of targeted nanoparticles with cells via ANN model could lead to a more rapid, more convenient and less expensive approach in comparison to numerous experimental variations.
Open Access
An evolutionary basic design tool
(2010) Akbulut, Dilek
As a creative act, design aims at achieving innovative solutions to fulfill the requirements provided in the problem definition. In recent years, computational methods began to be used not only in design presentation but also in solution generation. The study proposes a design methodology for a particular basic design problem on the concept of emphasis. The developed methodology generates solution alternatives by carrying out genetic operations used in evolutionary design. The generated alternatives are evaluated by an objective function comprising an artificial neural network. The creative potential of the methodology is appraised by comparing the outputs of test runs with the student works for the same design task. In doing so, three different groups of students with diverse backgrounds are used.
Open Access
Fire and flame detection methods in images and videos
(2010) Habiboğlu, Yusuf Hakan
In this thesis, automatic fire detection methods are studied in color domain, spatial domain and temporal domain. We first investigated fire and flame colors of pixels. Chromatic Model, Fisher’s linear discriminant, Gaussian mixture color model and artificial neural networks are implemented and tested for flame color modeling. For images a system that extracts patches and classifies them using textural features is proposed. Performance of this system is given according to different thresholds and different features. A real-time detection system that uses information in color, spatial and temporal domains is proposed for videos. This system, which is develop by modifying previously implemented systems, divides video into spatiotemporal blocks and uses features extracted from these blocks to detect fire.