Browsing by Author "Cengiz, Ramazan"

Now showing 1 - 1 of 1

Location Estimation of Base Station of Mobile Phones Using Artificial Neural Networks
(Kadir Has Üniversitesi, 2016) Cengiz, Ramazan; Öğrenci, Selçuk
Finding the location of the base station of mobile phones is a problem. There are different approaches and methods for finding location in the literature. A solution to this problem is using “supervised learning” method. Recently supervised learning technology is used in solving problems. For example computer is trained and fed by necessary data. This data includes input and ideal targets. After computer is trained properly it can make estimations of new data and give reasonable results. For the base station estimation program mobile phone data(as input) of mobile phones and base station locations(as targets) are given to software. Software can estimate base location with input data after training software Matlab is a good program for supervised learning methods and artificial neural networks. For finding the optimal neural network topology and optimal training methods Matlab is used in this project. Experiments are made with different datadifferent neural networks and different training methods. Results and outputs are observed. After finding best neural network topology experiments are made with real data. Raw data are collected from mobile phones and after converting it into useful numbers and after normalizing these numbers performance of the program is evaluated. Finally this solution is implemented in Java platform. in this thesis we present base location estimation program. This program takes as input a training data and test data. it uses training data to estimate the output of the test data. As output it creates a report that shows the estimated base station location.