13 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 10 / 13
Yayın On the extraction of the channel allocation information in spectrum pooling systems(IEEE, 2007-04) Öner, Mustafa Mengüç; Jondral, Friedrich K.The spectrum pooling strategy allows a license owner to share a part of his licensed spectrum with a secondary wireless system (the rental system, RS) during its idle times. The coexistence of two mobile systems on the same frequency band poses many new challenges, one of which is the reliable extraction of the channel allocation information (CAI), i.e. the channel occupation of the licensed system (LS). This paper presents a strategy for the extraction of the CAI based on exploiting the distinct cyclostationary characteristics of the LS and RS signals and demonstrates, via simulations, its application on a specific spectrum pooling scenario, where the LS is a GSM network and the RS is an OFDM based WLAN system.Yayın Mobile applications discovery: a subscriber-centric approach(Wiley Periodicals, 2011-03) Erman, Bilgehan; İnan, Ali; Nagarajan, Ramesh; Uzunalioğlu, HüseyinRapid adoption of smartphones and the business success of the Apple App Store have resulted in the rampant growth of mobile applications. Seeking new revenue opportunities from application development has created a gold rush. However, free or very cheap applications constitute a great bulk of the application downloads putting great pricing pressure on the developers. Furthermore, usage statistics suggest that most of the applications have been either one-trick applications or are downright useless, meriting no attention from the user beyond the first day. This is not surprising since cheap prices will dissuade developers from investing large sums of money to continue to develop more sophisticated, high quality applications. Developers have been complaining about the lack of visibility of their applications in stores that are beginning to resemble a high volume warehouse. It is clear that enhancing application discovery and building better marketing tools will be essential for the continued success of the mobile application marketplace and application stores. This paper proposes and investigates techniques for effective discovery of applications by matching user interests with application characteristics, with a special focus on adapting classical data mining techniques to user ratings of the applications. The user ratings are leveraged to make recommendations on potential applications of interest.Yayın Soft decision trees(IEEE, 2012) İrsoy, Ozan; Yıldız, Olcay Taner; Alpaydın, Ahmet İbrahim EthemWe discuss a novel decision tree architecture with soft decisions at the internal nodes where we choose both children with probabilities given by a sigmoid gating function. Our algorithm is incremental where new nodes are added when needed and parameters are learned using gradient-descent. We visualize the soft tree fit on a toy data set and then compare it with the canonical, hard decision tree over ten regression and classification data sets. Our proposed model has significantly higher accuracy using fewer nodes.Yayın Generative and discriminative methods using morphological information for sentence segmentation of Turkish(IEEE-INST Electrical Electronics Engineers Inc, 2009-07) Güz, Ümit; Favre, Benoit; Hakkani Tür, Dilek; Tür, GökhanThis paper presents novel methods for generative, discriminative, and hybrid sequence classification for segmentation of Turkish word sequences into sentences. In the literature, this task is generally solved using statistical models that take advantage of lexical information among others. However, Turkish has a productive morphology that generates a very large vocabulary, making the task much harder. In this paper, we introduce a new set of morphological features, extracted from words and their morphological analyses. We also extend the established method of hidden event language modeling (HELM) to factored hidden event language modeling (fHELM) to handle morphological information. In order to capture non-lexical information, we extract a set of prosodic features, which are mainly motivated from our previous work for other languages. We then employ discriminative classification techniques, boosting and conditional random fields (CRFs), combined with fHELM, for the task of Turkish sentence segmentation.Yayın Regularizing soft decision trees(Springer, 2013) Yıldız, Olcay Taner; Alpaydın, Ahmet İbrahim EthemRecently, we have proposed a new decision tree family called soft decision trees where a node chooses both its left and right children with different probabilities as given by a gating function, different from a hard decision node which chooses one of the two. In this paper, we extend the original algorithm by introducing local dimension reduction via L-1 and L-2 regularization for feature selection and smoother fitting. We compare our novel approach with the standard decision tree algorithms over 27 classification data sets. We see that both regularized versions have similar generalization ability with less complexity in terms of number of nodes, where L-2 seems to work slightly better than L-1.Yayın Parallel univariate decision trees(Elsevier B.V., 2007-05-01) Yıldız, Olcay Taner; Dikmen, OnurUnivariate decision tree algorithms are widely used in data mining because (i) they are easy to learn (ii) when trained they can be expressed in rule based manner. In several applications mainly including data mining, the dataset to be learned is very large. In those cases it is highly desirable to construct univariate decision trees in reasonable time. This may be accomplished by parallelizing univariate decision tree algorithms. In this paper, we first present two different univariate decision tree algorithms C4.5 and univariate linear discriminant tree. We show how to parallelize these algorithms in three ways: (i) feature based; (ii) node based; (iii) data based manners. Experimental results show that performance of the parallelizations highly depend on the dataset and the node based parallelization demonstrate good speedups.Yayın Milkminer: a dairy farm analysis and learning system(Işık Üniversitesi, 2014-01-23) Genç, Ayşenur; Ünel, Gülay; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Enformasyon Teknolojileri Yüksek Lisans ProgramıAgriculture and animal breeding industry is getting its share from the rapid advances in technology which enables the world wide use of automation systems. The use of automatic milking systems in dairy farms for milking has also increased. The goal of this thesis is to contribute to the development of performance enhancing processes by analyzing the data collected in dairy farms for discovering new rules and relationships using data mining. The outputs of this thesis will be used by Triodor Company as an analysis and learning system. Currently, the data collected in various dairy farms all around the world that use automated systems is stored for use in local or global databases as a part of the automation projects. Since these automation technologies are still in the development phase, research and development on detailed study, analysis and relationship recognition among data is in early stages and limited. Currently, the central database populated by the software technologies developed by Triodor contains data about each individual farm such as key performance measurements collected daily from more than 30 countries and 4000 dairy farms for various types of users. This large scaled dataset is not used for any operation or analysis other than simple query answering. If this large database and external data sources (such as weather, vegetation) are analyzed for the detection of potential hidden relationships among data then it will be possible to realize improvements in these farms regarding various types of criteria such as performance, sustainability, and product quality. In this thesis, an analysis and learning system that works on the data collected from dairy farms is developed. In the implementation of this system, the dairy farm database design is analyzed in terms of the content to be used in analysis purposes, and then the database is analyzed using data mining methods. As a result, a system with quantitative analysis techniques via data mining methods is generated.Yayın An experimental evaluation of prior polarities in sentiment lexicons(IEEE, 2017) Kanburoğlu, Ali Buğra; Solak, ErcanWe present the results of an experiment to assess the validity of prior polarities available in sentiment lexicons. We designed a ranking task that was elicited through pairwise comparisons and compared the results to those predicted by two popular sentiment lexicons. We find that the experiment results show a moderate level of agreement between the lexicons and human judgments.Yayın Forecasting and analysis of domestic solid waste generation in districts of istanbul with support vector regression(Institute of Electrical and Electronics Engineers Inc., 2020-10-12) Özçelik, Şuayb Talha; Tek, Faik BorayWaste planning is essential for large and developing cities such as Istanbul. In this report, we perform data analysis on "Waste Amount Based on District, Year and Waste Type"dataset shared by Istanbul Metropolitan Municipality. After analyzing the waste of the districts, we used support vector regression (SVR) to forecast the waste amounts for the coming years. The analysis has shown an overall increasing trend in the waste generation, although it dropped in 2019. The SVR predicts that the most waste generating district will be Küçükçekmece in the coming years.Yayın Data mining techniques customer relationship management : a case study for Doğuş Otomotiv(Işık Üniversitesi, 2003-10) Gören, Olga Ufuk; Yarman, Bekir Sıddık Binboğa; Işık Üniversitesi, Fen Bilimleri Enstitüsü, Enformasyon Teknolojileri Yüksek Lisans ProgramıCustomer Relationship Managment has recently become a very important investment for many companies. The value of the customer is expressed more clearly with this new concept and it's demonstrated through its applications. The increased importance of focusing on the customer has been recognized in related academic studies in recent years. CRM has great importance in helping companies find the answers to questions such as ''What types of customers exist in which markers?'', ''What do they demand from companies?'', ''Who should be targeted as a customer?''. In fact it is a corporate strategy itself. In order to be successful, a company needs to combine the whole CRM concept with its vision. Furthermore, CRM is a strategy which aims to increase the improtance of every stakeholder of the company. With CRM, it is easy to realize the great value sone data for future of a company even though such data for the future of a company even though such data were not considered important in the past. CRM also provides ways of evaluating these data, and creating new goals out of results. The aim of this thesis is to study the CRM implementation process and show how customer data is being classified, analyzed and assessed, making use of the results of RFM analysis and demographical analysis as implemented in Doğuş Otomotiv. The importance of this thesis is that it keeps customers and their information in a data warehouse, so as to be available for effective use and analysis when needed. This analysis will help companies organize their marketing and sales activisties so as to respond to the right customer, at the rigth time, in the right way, and at the right place within their market.
