MSc in Financial Economics
Bu koleksiyon için kalıcı URI
Güncel Gönderiler
Öğe The relationship between bitcoin, ethereum and nft market(İstanbul Bilgi Üniversitesi, 2022) Çakır, Ersel; Demir, EnderABSTRACT: The popularity of NFT tokens has started to spread to the masses as the NFT market has grown tremendously from the beginning of 2021. This study examines the relationship between the NFT market and Bitcoin and Ethereum through Chiliz, MANA and THETA. In the study, the relationship between the prices of Bitcoin, Ethereum prices and the prices of three different elements of the NFT market with high market capitalization is examined on the basis of interrelationships. Using daily data between January 2020 and June 2022, short-term dynamic Granger causality of Ethereum on Bitcoin, Granger causality of THETA on Ethereum and Granger causality of THETA on MANA are shown. However no effects of Bitcoin on Ethereum, MANA, Chiliz and THETA can be shown. The results suggest that THETA may have a causal effect both on Ethereum, which is the second largest actor in the cryptocurrency market, and on the MANA token, which is one of the important representatives of the NFT market. From this point of view, it can be said that THETA contains preliminary information about the NFT and cryptocurrency markets.Öğe Determinants of credit use of households in Turkey: the effect of Covid-19(İstanbul Bilgi Üniversitesi, 2022) Özden, Mehmet Rauf; Öztürk, Serda SelinABSTRACT: In this study, it is investigated that which factors determine the credit use of households and what the effects of Covid-19 period have been on the credit use of households in Turkey. Furthermore, it is also examined what factors influence the bank preference of Turkish households. The study is done with survey methodology. The relationship between the demographical features of attendees and the factors taken into consideration, factors that are influential while using credit and factors taken into consideration for the bank preference is analysed with one-way ANOVA and t-test via SPSS. On the other hand, the effects of factors taken into consideration, factors that are influential, factors for the bank preference and the effects of demographic and Covid-19 factors on the credit use of households are analysed with Structural Equation Model (SEM) via STATA. It’s seen that arising the need, having cash shortage and the willingness to close another debt or credit are more influential in the decision of females to use credit compared to men. Moreover, it’s found that females are more sensitive in terms of interest rates compared to men while choosing a bank. The number of children, the age and the monthly income level of individuals influence the interest free (Islamic) finance preference of households. The general outlook of the economy, expectation for the future of the economy and political developments are the most important factors that households take into consideration while using credit. For the factors that are influential when households use credit, having cash shortage is seen to be the most remarkable factor. Satisfaction with customer relations is found to be the most powerful factor on the bank preference of households. The number of children, the age and the marital status of individuals are the most influential demographic factors on the credit use of households, respectively. Lastly, it’s reached that Covid-19 period has significantly affected the credit use of households.Öğe For a fistful of dollars - effect of exchange rate shocks on the labor supply of online gig economy in developing countries(İstanbul Bilgi Üniversitesi, 2021) Demiröz, Ahmet Taha; Reis, EbruABSTRACT: In this study, the phenomenon of the gig economy (also called the sharing economy, platform economy, or on-demand economy) and its growth over the last 3 years is analyzed through the perspective of developing countries. The focus of this study is the online gig economy in developing countries and the factors affecting people to supply their labor on the online marketplace. Studies between unemployment and online gig economy labor supply have been conducted in the past, however, these studies have concerned online labor in developed countries (mainly the US) and therefore did not account for an important factor for people in the developing world supplying their labor in the online marketplaces: getting paid with stable foreign currencies. Therefore, in this study, using the data provided by Oxford Internet Institute's Online Labour Index, the last 3 year's worker supply data is empirically analyzed with countries that went through an exchange rate shock in the last 3 years. This study aims to shed more light on the factors that affect the workforce in developing countries to supply their labor on the online gig economy, by adding the variable of exchange rates (or rather weak currencies against the dollar), which has been absent from the literature on the gig economy thus far. The results of the study are inconclusive, because of data limitations. Obtaining worker supply data specifically for the research on the effects of exchange rate shocks on the online gig economy may provide healthier results between these two factors and the phenomenon of online outsourcing in general.Öğe Analysis of public perception of central banks digital currencies: A survey analysis(İstanbul Bilgi Üniversitesi, 2021) Pişkin, Filiz; Özyıldırım, CenktanABSTRACT: Human is the creature that can adapt most easily to all kinds of changes. Along with the technological revolutions experienced around the world, our habits and needs are changing rapidly. “Saving time” is one of the things that humanity needs most today. New approaches emerging both in sociological, economic and scientific dimensions stem from the need for rapid and excessive consumption. In return, it forces you to think fast and produce fast. On the basis of the concept of economy, which is the focus of production and consumption, money and its functions lie. Money, like all other social and economic elements, has begun to evolve with the digital transformation. So, it's time to look ahead and consider what kind of money and payment systems will be needed to meet the needs of the increasingly digital economy. With the transition to Digital Transformation and Digital Economy, paper banknotes, the most common form of money, began to give way to payments with debit cards, virtual POS and QR codes. This evolution in the economy gained different dimensions with the inclusion of the Blockchain Algorithm in payment systems. The use of the Distributed Ledger structure together with the concept of crypto assets has created a new formation such as decentralized digital money, which cannot yet be classified as a type of money. Today, a decentralized payment system contains many uncertainties that are not reliable in terms of the economic policies of the States. Therefore, as a result of the development of the Digital Economy, Central Banks print their Central Bank Digital Money (CBDC) electronically and present them to the public as a new payment system through digital wallets. Some countries create their own CBDC systems, while others try to adapt applied models. Therefore, it is very important to understand the current models in terms of adapting CBDC in the form of National Official Digital Currency. In the literature review, the reasons for many countries taking action regarding CBDCs; The emergence, development and results of the Bahamas Sand Dollar Project, which is accepted as the first official retail CBDC example and officially started to be implemented in 2019, have been seen to be effective. In addition, in the Working Paper published by the Bank of England in 2020, the xiii opportunities and challenges that CBDCs will create when they are implemented in the UK are discussed with their most critical points in terms of the risks they may create in financial markets, society and institutions through the Platform Model. These two models were also taken into account when other countries were creating their transitional designs to CBDC. In the continuation of this study, the studies of the main countries that have made progress on CBDCs in the world are mentioned too and information about the current situation is given. With these inferences from the models, it has been determined that CBDCs have many important advantages such as tax evasion, preventing financial terrorism such as fraud and providing low cost, but they have disadvantages that can negatively affect monetary policy, financial stability and the banking sector if not implemented well. A survey study was conducted on the online platform to measure the public perception of the advantages and disadvantages determined. In this survey study, 3 questions about gender, age and the sectors they work for people who have a financial background and work in the finance sector and those who do not work in the finance sector; A total of 26 questions were asked, 23 of which were Likert scale questions, about the advantages and disadvantages of CBDC. A total of 317 people were reached through the survey. The data collected as a result of the research were evaluated with Cronbach's Alpha and Independent Sample t-Test, and it was observed that people with a financial background were significantly different from those who are not from financial sector, and that the public did not have enough information about CBDCs according to the sample group. So, Central Banks should be very careful when transitioning to Digital Currencies and raise public awareness on this issue.Öğe The effect of FDI on developing countries(İstanbul Bilgi Üniversitesi, 2021) Akyol, Yusuf İbrahim; Soybilgen, BarışABSTRACT: With the development of the globalization process, capital transfer between countries has become very important. In this context, it is seen that developing countries mostly focus on foreign direct investments in terms of their economic income. The liberalization trend in the 1980s, caused capital mobility has grown rapidly and widely. In this process, foreign direct investments started to flow to developing countries, firstly to benefit from the natural resources of these countries and to benefit from cheap labor, and then to be partner of market shares of these countries. After these developments, the real impact of foreign direct investments on the economic growth of host countries has been discussed continuously. Although there is no concurrence on the relationship between foreign direct investment and economic growth in the economics literature, the increasing view in recent years shows that foreign direct investments contribute to economic growth. This thesis investigates the effect of foreign direct investments on the economic growth of 29 developing countries during the period of 1991- 2017 by using the panel data analysis method.Öğe Gold investment attitudes of individuals-Turkey example(İstanbul Bilgi Üniversitesi, 2021) Kılıç, Sevcan Öztürk; Öztürk, Serda SelinABSTRACT: Throughout history, the gold has been a commonly-used precious metal demanded for an instrument of payment, a substance for jewelry manufacture, an industrial output and particularly an investment tool in recent years. Both the earmarks of gold itself and the investor profile play a crucial role in the gold investment attitudes. Starting from the point of view that not only rationality but genetic, psychological, demographical, or sociological facts bias individuals’ investing decisions, in this study, the tendency of individual investors in Turkey while investing in gold and the reasons that they consider while investing in gold will be presented by survey method and examining the relations between their perceptions and preferences with their demographics by one-way ANOVA and t-test and whether there is a relation between the perceived benefits and demographical traits of individuals and investment behavior and strategies of individuals via Structural Equation Modeling.Öğe Analysis of the relationship between cryptocurrencies and Borsa Istanbul: Before and after COVID-19(İstanbul Bilgi Üniversitesi, 2021) Karadeniz, Önder; Öztürk, Serda SelinABSTRACT: Today, cryptocurrencies are now used as an investment tool and it is thought that the crypto money sector will find investors in the future. Bitcoin, which is the oldest currency among cryptocurrencies, is 12 years old and is an investment tool accepted in the world market. In this study, it is aimed to examine the effects of fluctuations in cryptocurrencies on Borsa Istanbul Bist100 index. The sample of the study included the Bist100 Index, Bitcoin, Ethereum, Ripple, Litecoin, Monero and Iota data within Borsa Istanbul in 17-12-2018 between 17-12-2020. There are total of 499 observations. This study is divided into 2 periods as pre-pandemic and post-pandemic. The reason for the study to be divided into 2 periods is how much the pandemic has affected the crypto money market and the Bist100 index. In order to reach a healthier result, the logarithm of the studied data was taken and the logarithmic return was evaluated. In this study, Dynamic Conditional Correlation GARCH model was used. The focus of the study is "volatility". GARC & ARCH analyzes were performed on the financial data used for the 2018-2021 period of the data selected as a sample in the analysis. Before the DCC-GARCH analysis, the Augmented-Dickey Fuller test applied to measure the stationary and ARCH LM test were performed to check whether there was an ARCH effect in the series. As a result of the study, it was concluded that the volatility between Bitcoin and Bist100 variables affect each other widely over time.Öğe Determinants of life and non-life insurance demand: Evidence from OECD countries(İstanbul Bilgi Üniversitesi, 2021) Şimşek, Nevzat Fatih; Demir, EnderABSTRACT: Insurance supports the economy and contributes to investments as a precaution against risks. Therefore, developing countries should determine appropriate insurance policies in order to support their economic structures. As it is known that the insurance sector has positive contributions to the growth of countries, it has an important place in the economic development of countries. In this article, four models have created for the penetration and density of life and non-life insurances in The Organization for Economic Development and Cooperation (OECD) countries consisting of 36 countries, taking into account the annual data between 1996 and 2017, to measure the effects of economic variables on the insurance development of countries and tested with panel data analysis method. The research aims to investigate the determinants of the consumption of life and non-life insurances in OECD countries and to identify the economic impact of these determinants. Thus, the insurance sector and indirectly the economic development moves can be determined. The study results found that while the GDP per capita and economic freedom increased the penetration of life insurance, consumer price inflation decreased this rate. Foreign direct investment and age dependency ratio negatively affect non-life insurance penetration, while per capita GDP, urban population, and economic freedom have a positive effect. While foreign direct investment and economic freedom increase the density of life insurance, consumer price inflation and the urban population ratio cause a decrease. It has been observed that the dependency ratio does not carry a statistically significant burden on the life insurance side. For non-life insurance density, the age dependency ratio and the urban ratio has a negative effect whereas life expectancy, foreign direct investment, and economic freedom have a positive impact. Consumer price inflation does not seem to have a statistically significant effect on non-life insurance demand in OECD countries.Öğe Evaluation of credit risk impact on bank profits in volatile periods: case of 2008 subprime Mortgage and 2012 European financial crises(İstanbul Bilgi Üniversitesi, 2021) Özçetin, Merve Aysun; Öztürk, Serda SelinABSTRACT: This dissertation is an inquiry into the credit risk impact on bank’s profitability ratios, which are Return on Asset (ROA) ratios by paying attention to volatile periods focusing on the time period between 2008 Subprime and 2012 European Financial Crises. The main two questions are how the bank’s profitability ratios are affected by the global financial crises and are these ratios of banks in developed countries less affected by global financial crises. The data samples were collected from 40 banks covering the years from 2008 to 2013. Banks included in the analysis are classified as developed and underdeveloped, considering the country classification defined by the World Bank for the year 2008. In this research, the panel data analysis is used since the dataset of randomly selected banks has both time series and cross-section dimensions with the fiction of the time and a variety of banks classified as developed and underdeveloped countries. Return on Asset ratio was used as a dependent variable, and as explanatory variables are defined measurement of risk, Non-performing Loans to Total Loans, Tier 1 Capital Ratio, Total Loans to Total Deposits variables were used. The results indicated that in the examined period with the sample banks, Non-performing Loans to Total Loans variable showed a greater risk for banks located in underdeveloped countries than banks in developed countries. Tier 1 Capital Ratio, a bank’s core capital measurement, has much greater significance as a risk reduction in banks in underdeveloped countries. Banks in underdeveloped countries have much greater risk than banks in developed ones for Total Loans to Total Deposits, since an increase in the total loan amount and a decrease in the total deposit amount mean an increase in the risk in the banking system. When the effect of explanatory variables on banks’ profitability ratios is examined, the risk of banks in underdeveloped countries is higher than the risk of banks in developed countries in the stated period of time.Öğe The impact of financial development on economic growth: Evidence from OECD countries(İstanbul Bilgi Üniversitesi, 2020) Şendoğan, Gülseda; Öztürk, Serda SelinABSTRACT: Financial development is a crucial factor for all countries because it directly affects economic development, which means that it affects every single unit of a country. More than that, financial development is the pillar of a country's independence. This is why it encourages economic growth in terms of capital appreciation, technological progress, production, and investment. Thus, the question of how increasing financial developments in the post-globalization period and, which channels affect economic growth has been a topic that has been explored for a long time. While empirical studies often establish a direct link between indicators that provide financial development and growth, discussions continue on how these results should be interpreted. This study tries to shed light on this relationship with OECD countries. While doing this, the panel data analysis method was used, and the findings obtained as a result of the study show the existence of a significant relationship.Öğe Comparison of the performances of BIST 30 portfolios by using machine learning algorithms(İstanbul Bilgi Üniversitesi, 2021) Erik, Fatih; Öztürk, Serda SelinABSTRACT: In this study, the prices of BIST 30 stocks are estimated using machine learning algorithms such as linear regression, decision tree, support vector machines, long- short term memory and XGBoost, and based on these predicted prices, various portfolios have been created. Portfolios are generally managed in two ways, active and passive. The most important factor while choosing one is their return. Current studies show that the returns of funds traded on Borsa Istanbul Stock Exchange performed below the market, in other words, passive management beat active management. On the other hand, there are also studies showing that portfolios to be created with certain strategies yielded higher returns than the market. In this study, first of all, the prices of BIST 30 stocks for 1 month, 2 months, 3 months, 4 months, 6 months and 12 months are predicted by using machine learning algorithms and according to these revision periods, 6 strategies were created. In the second stage, these strategies were diversified by changing the portfolio sizes (5-10-15 shares). As a result, a total of 90 strategies were formed with 3 different sizes, 6 different timeframes and 5 different algorithms. As a result of the study, it was determined that the returns of portfolios created by using machine learning algorithms are generally above the return of the BIST 30 index. Also, the LSTM algorithm generally makes more successful predictions, 12-month strategies yield higher returns than other strategies, and portfolios including 5 shares are more successful than other portfolios. The result table shows that the most successful portfolio is the portfolio with 5 shares, revised every 12 month and created by using the XGBoost algorithm.Öğe Economic impact of a highway construction project(İstanbul Bilgi Üniversitesi, 2021) Bayındır, Emre; Öztürk, Serda SelinABSTRACT: Economic impacts of infrastructure projects have been researched and analyzed through different methods. Highways and state roads recently constructed contribute to economic development on several counts in Turkey. The subject of this study was to estimate the economic impacts of any infrastructure investment on the entire economy by using input-output analysis. For this purpose, it aimed to determine the total impacts of a selected highway project on the Turkish economy during the construction phase of a 5-year period as a case study. The project, whose economic effects will be determined is specific to 100 km in a region and it is on one of the main arterial roads in Turkey. The input-output model was applied to calculate the direct, indirect, and induced impacts. In the study, national input-output tables created by TurkStad in 2012 has been used and total output or requirements, gross value-added and employment multipliers concerning highway construction in Turkey were computed separately to calculate relative changes in the economy for grand total and each year. The findings showed that the construction of selected highway has obviously positive impacts on the national economy in terms of output, gross value added and employment. Thus, this applied method can be a good approach to quantify economic effects on both regional and national economy for policymakers and used for other infrastructure projects.Öğe Predicting government bond price returns using machine learning algorithms(İstanbul Bilgi Üniversitesi, 2021) Özdemirci, Deniz; Soybilgen, BarışABSTRACT: In this study, one-step, five-step, ten-step, thirty-step, sixty-step and ninety-step ahead daily close price returns of selected common long-end term bonds of Turkey, Germany and the United States are predicted with machine learning models, which namely are the linear regression, the random forest, gradient boosted regressor trees and support vector regressors. This study also provides information about how bonds’ unique security codes (more specifically, how the National Security Identification Numbers and the International Security Identification Numbers) are determined, how the bond market and public debt evolved within the date range of the data set and the reasoning of the selection of the selected independent variables is provided: Explanatory variables are selected among other price series that are expected to be relevant with the dependent variable bonds. Selected performance metrics to evaluate each model and compare them with the baseline model, the random walk model, are the “Residual Sum of Squared Error”, “Mean Absolute Percentage Error”, and the directional accuracy. Results suggest that with the selected combination of models and independent variables, the gradient boosted regressor model and the linear regression model outperform other used machine learning models, while all models outperform the indicated baseline model, the random walk.Öğe The determinants of CDS: an emerging markets analysis(İstanbul Bilgi Üniversitesi, 2021) Yıldız, Serhat; Öztürk, Serda SelinABSTRACT: In this study we tried to analyze the relationship between CDS spreads and other financial and macroeconomic indicators such as stock index, foreign exchange rate, industrial production index and consumer price index. We focused on emerging markets, especially BRIC countries and Turkey. We applied unit root tests, cointegration tests, vector error correction models, vector autoregressive models and Granger causality tests. Our results imply that the relationship between CDS spreads and other financial variables and macroeconomic variables are not homogeneous across countries. Nevertheless, a strong implication of the results is that the causality generally runs from financial variables, in particular stock index to CDS spreads.Öğe Estimation of static and dynamic optimal hedge ratios: an application to the bist30 index futures(İstanbul Bilgi Üniversitesi, 2021) Uçar, Selin; Öztürk, Serda SelinABSTRACT: Especially in recent years, due to fluctuations in both domestic and international financial market, investors pay more attention to control the risks exposed due to their positions in the markets. Various econometric methods have been developed to estimate the optimal hedge ratio. In the literature, some of researchers studied with the static models, while others used dynamics models for estimation of the optimal hedge ratio. The aim of this paper is determining the optimal hedge ratio by providing a comparision of various econometric models. The daily closing prices of BIST 30 Index and BIST30 Index Futures are used for forecasting and the optimal hedge ratio are estimated by employing static models such as the Ordinary Least Squared (OLS) model, the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models , the Error Correction Model (ECM) and the Vector Error Correction Model (VECM). Also we employ the bivariate VECMDiag- BEKK-GARCH model to estimate dynamic hedge ratio. The hedging performance is measured in terms of the variance reduction provided by each models. According to the findings, the VECM and the GARCH models provide only slightly better performance than the traditional linear regression model and the bivariate VECM-Diag-BEKK-GARCH model fails to outperformed the static models.Öğe Gold market in Turkey and effect of pandemic on gold(İstanbul Bilgi Üniversitesi, 2021) Dikmen, Okan; Öztürk, Serda SelinABSTRACT: The Covid-19 epidemic, which can be called the plague of our time today, affects financial markets as deeply as it affects the health sector. Analysis in a study to measure the impact of the Covid-19 pandemic on gold prices, 22.01.2020 -08.04.2020 was estimated using the ARDL model with daily data. In the study, which examined gold prices as dependent variables, Covid-19 Turkey case number, Covid-19 World Case Number, US dollar ratio, policy interest and fuel prices were included in the analysis as independent variables. Because the selected variables are stable at different levels, the results of short-term and long-term analysis were interpreted within the framework of the ARDL model. Based on the results of the error correction model, the long-term capture rate of the model was estimated at 51 percent. Gold is the most well-known and traditional financial instrument in which we can observe how investors develop financial instruments in situations that lead to global panic. According to the results, there are negative correlation between gold prices and short-term Unites States dollar exchange rate and political interest rate. Also, there are positive correlation between gold prices and fuel prices, Covid-19 World number of cases, Covid-19 Turkey number of cases. According to the results of a long-term relationship Covid-19 has been observed that a co-integrated relationship between gold prices with all variables except for the case of Turkey the number of variables.Öğe Users’ perception and continued usage of m-banking(İstanbul Bilgi Üniversitesi, 2021) Yurdakul, Murat; Öztürk, Serda SelinABSTRACT: In this study, the determinants affecting the continuous usage of mobile banking were examined. Motivated by the need to understand the variables that success, sustenance, and long-term development of mobile banking depend on. The proposed research model is developed by the incorporation of Technology Acceptance Model, Task Technology Fit, and perceived risk into Expectation Confirmation Model and Structural Equation Modeling is used to test the research model with empirical data gathered by a survey from 154 mobile banking users who use m-banking services offered by banks in Turkey. The results show that continuance intention is affected by perceived task technology fit, perceived usefulness, perceived risk, and satisfaction, significantly. But the effect of perceived ease of use on continuance intention is insignificant. Besides, as the main predictors, perceived task technology fit, and expectation confirmation affect satisfaction. Moreover, perceived usefulness is determined by perceived task technology fit, perceived ease of use, and expectation confirmation. The results also indicate that gender significantly moderates the effect of perceived usefulness to continuance intention.Öğe Modelling minimum hurdle ratio for financial sector pricing in emerging market(İstanbul Bilgi Üniversitesi, 2021) Sargan, Kemal Çağdaş; Öztürk, Serda SelinABSTRACT: One of the most crucial rules for achieving a sustainable and stable growth in financial markets is the development of correct pricing models. Especially in highly volatile markets, firms need to develop advanced pricing models in order to acquire expected return and achieve the sustainable growth trends. Finance sectors in every country are interested in the pricing models more than others. Fund Transfer Pricing (FTP) method for financial institutions has started to be used already, and as a pricing process that increases the importance emerges through the day. FTP model is very efficient method to handle liquidity, interest and option risk for companies, and researchers are focusing to improve this model by developing additional behavioral models (pre payment models, etc.). In particular, it was once again understood how important it is to include risk factors in these strategies after the financial crisis in America. In this study, it is aimed to improve the advanced FTP model with additional factors and to include credit risk models more effectively in the Hurdle model. In the first part, basis interest, liquidity Premium, and other variables that may affect pricing are described, and the second part is the modeling of credit risk factor is studied. Especially the contribution of behavioral and credit risk adjustments to the FTP model is one of the most important outputs of the study. The work will make a significant contribution to banks and financial institutions in emerging markets to manage their risks more efficiently and to create correct pricing strategies.Öğe Quantitative value investing approach evidence from Turkish stock market(İstanbul Bilgi Üniversitesi, 2021) Tüzüner, Emre Cem; İkizlerli, DenizABSTRACT: This thesis constructs an actively managed portfolio using a quantitative investing approach on selected publicly listed stocks in Istanbul Stock Exchange. After looking at available literature on alpha generating portfolio strategies, the thesis builds a new investment strategy based on stock selection criteria which includes several operational filters applied to all stock universe between 2010 and 2020. The regression results prove that this investment technique is an alpha generating strategy which beats the benchmark index’s performance as well as other comparable Turkish funds’ performance for the given period.Öğe Income estimation model for individual customers(İstanbul Bilgi Üniversitesi, 2021) Sargan, Çağrı; Özyıldırım, CenktanABSTRACT: The aim of this thesis is to show the increase in predictive power of the income estimation model used during individual product allocation in financial institutions by using machine learning modeling techniques. There are hundreds of factors that affect a customer's income. Although most of these factors are the customer's own information, macroeconomic indicators can cause an impact on individuals' income. In the literature, generally traditional modeling techniques have been used to estimate the income of the customers, and in this study, a modeling study has been carried out by using boosting and bagging algorithms. Compared to regression-based modeling performances, it has been observed that the performance of boosting-based models has more explanatory power. With this study, it is aimed to create a more accurate revenue estimation mechanism for customers. In this way, credit limits will be defined to customers in direct proportion to their ability to pay, and default rates in the portfolio will be minimized with correct product allocation. Within the scope of the study, model validation tests were performed and it was determined that the model performance for the validation sample provided the most descriptive results with the XGBoost algorithm.
