3 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 3 / 3
Yayın Big data storage and automated text summarization in Turkish text(Işık Üniversitesi, 2018-06-19) Aysu, Erdinç; Yıldız, Olcay Taner; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Bilgisayar Mühendisliği Yüksek Lisans ProgramıThe subject of this study is storing the large datasets in accordance with Big Data ecosystem and to extract the summary sentences of a text in Turkish, apply the automatic text summarization process which is a subtopic of Natural language processing (NLP). For this purpose, Turkish news articles were collected and the study was carried out through these texts. For the performance test of the work done, 50 different news textiles were given to 20 different persons and 3 sentences which were considered important from each other were asked to be selected and their results were compared with each other. Then, the results from the people were compared with the results from this study. As a result of the test process, the summation performance of the work was measured approximately as thirty-six percentage.Yayın Driver recognition and driver verification using data mining technigues(Işık Üniversitesi, 2007-09-25) Benli, Kristin Surpuhi; Eskil, Mustafa Taner; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Bilgisayar Mühendisliği Yüksek Lisans ProgramıIn this thesis we present our research in driver recognition and driver verification. The goal of this study is to investigate the affect of different classifier fusion techniques on the performance of driver recognition and driver verification. We are using five different driving behavior signals for identifying the driver identities. Driving features were extracted from these signals and Gaussian Mixture Models were used for modeling the driver behavior. Gaussian Mixture Model training was performed using the well-known EM algorithm. In recognition study posterior probabilities of identities called scores were obtained with the given test data. These scores were combined using different fixed and trainable (adaptive) combination methods. In verification study we compared posterior probabilities with fixed threshold values for each classifier. For different thresholds, false-accept rate versus falsereject rate was plotted using the receiver operating characteristics curve. We observed lower error rates when we used trainable combiners. We conclude that combined multi-modal signal or classifier methods are very successful in biometric recognition and verification of a person in a car environment.Yayın Differentially private attribute selection for classification(Işık Üniversitesi, 2015-06-18) Var, Esra; İnan, Ali; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Bilgisayar Mühendisliği Yüksek Lisans ProgramıAny study on processing or analyzing large data sets that contain personally sensitive data should conform against some form of privacy protection mechanism. Otherwise, malicious people can aceess these data sets to extract private information and use this private information in agency operations, blackmail, fraud or any other harmful actions. Importance and necessity of privacy preserving data mining is increasing day by day, hence public and government lawmakers, privacy advocates and the media are drawing more and more attention to this subject daily. This thesis proposes an approach to that selects features from a data set according to the differential privacy mechanism and implements this proposed solution on a popular data mining library called WEKA.
