Keywords: Deep learning, Data processing, Deep neural network, Convolutional neural network, Auto encoder, Database

Data Processing and Analysis Using Deep Learning Approac

 Nowadays, information technology causes low cost production of powerful computers which lead to gathering, transferring, combining and storing the large amount of data. Increasing the size of databases leads organizations to extract information from storage data. However, data stored in databases cannot be used like a precious treasure for exploiting hidden information.

 In many cases humans are not able to distinguish and extract hidden information in the large scale data, so they need learning algorithm using database to help them. Therefore, in this thesis, some of the most important and widely used deep learning methods such as convolutional neural network, auto encoders, deep belief network, deep neural network, and recursive neural network are presented, then analyzing advantages and disadvantages of each one and comparing learning algorithm to each other, solutions to choose the most optimized data processing method and information exploitation are presented according to the database.

 Suggesting method of this study is the first deep learning method to achieve the best.() results in four areas of segmentation and interpretation of complex images, audio signal.() processing, text analysis and received signal analysis by six-antenna radar