Yüksek Lisans Tezleri

Permanent URI for this collectionhttps://hdl.handle.net/20.500.11779/1785

Browse

Filters

2017 - 20195

Settings

Department: search.filters.department.Lisansüstü Eğitim Enstitüsü, Büyük Veri Analitiği Yüksek Lisans ProgramıSubject: search.filters.subject.Logistic Regression

Search Results

Now showing 1 - 5 of 5
  • Master Term Project
    Alternative Credit Scoring Model for Thin File Customers
    (MEF Üniversitesi, Fen Bilimleri Enstitüsü, 2019) Korkmaz, İstem Akça; Taş Küten, Duygu
    Credit scoring is a widely used tool for banks, financial institutions or corporations. Traditional credit score models are calculated from past financial history of users, and this may lead to exclude some people who have limited financial history from the credit system. Alternative credit scoring allows sector players to access to a larger portion of these customers. The credit scoring industry has expanded with an "all data is credit data" approach that combines traditional credit scoring systems with new data points. In this study, we aim to build an alternative credit scoring model for customers who have limited financial historical data (thin file) by using alternative data points for a national bank in Turkey. Some of the alternative data points and variables have been gathered from one of the bank’s products: the authorized card for Turkish national league football tickets (Passolig). Using alternative data points combining with demographical and geographical information, we perform a comparison between the machine-learning approaches. We use logistic regression approach as a base model and perform a comparison between tree-based approaches: decision tree, random forest and XGBoost to select the most effective modelling approach
  • Master Term Project
    Software Projects Clustering and Selection by Machine Learning Methods
    (MEF Üniversitesi, Fen Bilimleri Enstitüsü, 2018) Torun, Elif; Ağralı, Semra
    In today’s hyper volatile business world, software development projects play key roles in maintain the current situation of the company and they are vital in taking the company one step further. Selecting the right project to invest is a critical decision point regarding the hard competition, diminishing profitability and high cost of the projects. The main aim of this study is clustering the projects and deciding which project to invest by using machine learning methods. We use IT project demands data of one of the biggest banks due to the capital, number of transactions and number of customer portfolio in Turkey. The data includes 2048 Information Technology related project demands occurred in 2017 and 2018. For the clustering part of the project both unsupervised and supervised learning methods are used and success rates are compared. We observe that supervised learning methods are more successful than the unsupervised ones. For the project selection part all process of the bank and output of the all steps are reviewed. According to our results, second workshop, which is the last step of the project assessment and selection process, has almost 50% of the total process effort and gives the precise effort estimation as an outcome, can be eliminated, and the project selection decision can be made with around 90% success ratio with machine learning methods. The result of this study provides an efficient way to select projects and a platform to see the complexity of the project portfolio.
  • Master Term Project
    Sms Spam Detection in Turkish Language
    (MEF Üniversitesi, Fen Bilimleri Enstitüsü, 2018) Gürkan, Cem Kaya; Koç, Utku
    Short message (SMS) is one of the most common communication methods. The growth of mobile phone users has led to a dramatic increase in using short messages. With the increasing number of mobile phone users, mobile phone users have started receiving unsolicited text messages. The use of SMS as a spam tool after the e-mail is due to a direct access to customer and high reversion to the users. These unsolicited short messages are disturbing the users even content intended for deceiving or defrauding (phishing). Up to date, all of the research carried out on SMS Spam detection was focused on the English language. In this study, Turkish datasets tagged with spam information is introduced and existing methods for English are applied to these datasets. The SMS dataset used in this study is gathered from different people and all messages are tagged according to whether they are spam or not. Naïve Bayes, Logistic Regression, SGD, SVM and Random Forest classification algorithms are tested with three feature extraction methods and a number of performance measures are evaluated. The evaluation resulted in a f-measure of 96.4% for SVM classification algorithm with TF-IDF (Term Frequency-Inverse Document Frequency) extraction method.
  • Master Term Project
    Trangling Weratedogs Twitter Data To Create Interesting and Trustworthy Explosatory/Predictive Anaylses and Visulation Using Different Machine Learning Algorithms
    (MEF Üniversitesi, Fen Bilimleri Enstitüsü, 2018) Arı, Esra; Çakar, Tuna
    Social media usage has rapidly grown in recent years and knowledge in these environments increased due to this expansion. Therefore, doing exploratory and predictive analysis from intensive data of social media became so popular. However, almost all of the large datasets obtained are uncleaned / raw data. Therefore, the assessing and cleaning of the data is at least as important as the exploratory and predictive analysis. The open source WeRateDogs twitter account tweets have been gathered, assessed, cleaned, analyzed and predicted for this thesis. As a result of the study, it was understood that the most important and most time-consuming part of the predictive data analysis is the data gathering and cleaning. As a result of this project, probability of dog’s breed whether retriever or not is predicted from the tweet’s text body. 24 points increase (%34 change) in accuracy values has been achieved by doing oversampling in the data sets which contain low event observation. At the same time, the decision tree, logistic regression and random forest algorithms are compared and it is shown that the random forest's model performance is better than the others. The algorithm works 13 points better than logistic regression, 21 points better than decision tree.
  • Master Term Project
    Fraud Detection In the Bitcoin Exchange Market
    (MEF Üniversitesi, Fen Bilimleri Enstitüsü, 2017) Namlı, Hüseyin; Güntay, Levent
    The trading volume and financial assets of Bitcoin are growing up, while the popularity of Bitcoin world increasing continuously in recent years. In parallel, the market becomes an attraction center for malicious people.