Emlak vergisi için CBS ve makine öğrenmesi teknikleri kullanılarak toplu taşınmaz değerleme modeli tasarımı

Designing a mass real estate valuation model using GIS and machine learning techniques for property taxation

PDF İndir

Tez No: 816348
Yazar: HALUK GÜL
Danışmanlar: DR. ÖĞR. ÜYESİ MAHMUT OĞUZ SELBESOĞLU
Tez Türü: Yüksek Lisans
Konular: Jeodezi ve Fotogrametri, Geodesy and Photogrammetry
Anahtar Kelimeler: Belirtilmemiş.
Yıl: 2023
Dil: Türkçe
Üniversite: İstanbul Teknik Üniversitesi
Enstitü: Lisansüstü Eğitim Enstitüsü
Ana Bilim Dalı: Geomatik Mühendisliği Ana Bilim Dalı
Bilim Dalı: Geomatik Bilim Dalı
Sayfa Sayısı: 97

Özet

Küresel ekonomik kaynakların önemli bir bölümünü oluşturan taşınmazlar, mülkiyete dayalı bir çok uygulamanın ana unsurudur. Bu sebeple taşınmaz değerinin, öznel yargılardan bağımsız, güncel yaklaşımlarla belirlenmesi oldukça önemlidir. Taşınmaz piyasası, bağlı olduğu değişkenlerle bulunan ilişkisi sebebiyle bir çok dalgalanmaya maruz kalabilmektedir. Bu dalgalanmaların yaşandığı piyasada taşınmaz değerinin geleneksel yöntemler kullanılarak belirlenmesi hukuksal ve ekonomik anlamda çeşitli sorunlara yol açmaktadır. Özellikle konusu bakımından taşınmaz değerine doğrudan bağlı olan Emlak Vergisi, vergiye esas rayiç değerin piyasa değerinden önemli ölçüde düşük kalmasından dolayı tartışılmaktadır. Bu bağlamda, ülkemizde, yasal mevzuat bütünlüğü çerçevesinde, uluslararası değerleme standartlara uygun bilimsel yaklaşımlara dayanan dinamik bir taşınmaz değerleme sistemine ihtiyaç vardır. Taşınmaz sayısı ve işlem hacmindeki artış göz önüne alındığında, taşınmaz değerlerinin daha hızlı ve yüksek doğrulukta sonuçlar veren toplu değerleme yöntemleriyle belirlenmesi için yeni yaklaşımlar geliştirilmiştir. Taşınmaz değerine etki eden konumsal faktörlerin Coğrafi Bilgi Sistemleri(CBS) ile analizi mümkün olmakta, makine öğrenmesi tekniklerinin kullanılmasıyla çok sayıda verinin analizi yapılmaktadır. Tez kapsamında öncelikle ülkemizdeki mevcut emlak vergisi mevzuatının analizi yapılarak tarihsel gelişimi, güncel işleyişi ve değer tespitinde yaşanan sorunlar ele alınmış, toplu taşınmaz değerleme konusu emlak vergisi çerçevesinde değerlendirilmiştir. Daha sonra makine öğrenmesinin türleri ve işleyişi incelenmiş, taşınmaz değerlemede yaygın olarak kullanılan makine öğrenmesi teknikleri irdelenmiştir. Çalışmada, taşınmazların toplu değerlemesi için CBS ve makine öğrenmesi tekniklerinin birlikte kullanıldığı kavramsal model tasarımı yapılmıştır. Kullanılacak model için CBS ortamında yakınlık, yüzey ve görünürlük olmak üzere örnek konumsal analizler yapılmış, analiz sonucu oluşan veriler değerleme için kullanılan faktörlerin ağırlıklarına göre birleştirilerek değer haritası üretilmiştir. Son olarak model kapsamında, makine öğrenmesi teknikleriyle yapılan analiz aşamalarında kullanılan araçlar ve yöntemler belirlenmiştir.

Özet (Çeviri)

Real estate is one of the fundamental elements that define the concept of property. When it is considered that the relationship between property and humans is closely linked, having real estate property has been seen as an important need and security. Although the use of property and the relationship between humans and property have changed over time, the concept of property has always maintained its importance. The increasing importance of the concept of property has led to the creation of the real estate sector, which is formed by land acquisition and increasing housing needs. The real estate sector is the foundation of financial investments in many countries and is sustained due to its perceived reliability and high potential for returns. The most important aspect of the real estate sector is real estate value. Determining real estate values has been a problem that has been worked on for a long time. The proper determination and maintenance of real estate values is important not only for real estate owners and real estate companies, but also for local governments that must define taxes based on the values of real estate and for agencies that will determine values for expropriation processes. The concept of real estate valuation has been a topic that concerns many disciplines. It is studied in a multi-disciplinary structure that concerns not only economics, geography, and city planning, but also business, finance, and statistics. Real estate valuation, which is influenced by developments in all these branches, is dynamic in nature. Therefore, the value of a real estate can be expressed as the final result that is determined transparently and impartially after a comprehensive study, taking into account the current effects of factors related to these branches. In Türkiye, the lack of comprehensive legislation regulating real estate valuation activities has caused many legal and technical deficiencies and problems. One of the unsolved problems in real estate valuation is the lack of a standard for valuation studies carried out by different institutions and organizations for different purposes. The formation of their own valuation committees by each institution, the lack of experience of the people serving in these committees in real estate valuation, and the scientific and technical inadequacy of the methods they use for real estate valuation are still relevant problems. For example, the real estate value subject to real estate tax, the value based on expropriation, and the value determined by experts appointed by the court, and the values given by bank experts for mortgages differ from each other, and the legal consequences arising from this difference create a significant workload and time loss for institutions. Within the current legal regulations, real estate tax is closely related to real estate valuation. The lack of a sustainable system for determining the real estate values that serve as the basis for the tax base and for tax collection indirectly causes many legal and economic problems. In the current legal framework in Türkiye, the values resulting from the studies on determining real estate value fall below the values in free market conditions, which reduces the revenue obtained from taxation and creates unfair results in taxation. In recent years, real estate valuation has begun to evolve from classical methods to mass valuation models. While real estate valuation using classical methods can be subjective and inconsistent, it takes time to determine the value of a large number of real estate properties using classical methods. To solve this problem, mass valuation models that use various statistical methods have been developed. Using objective value criteria that affect real estate value, comprehensive and dynamic real estate valuation models can be created through various analyses and value maps prepared on a large scale. With the widespread use of information technology and developments in the field of computing, the number and diversity of data available on various platforms has increased significantly. Artificial intelligence studies allow the analysis, classification, prediction, and clustering of these data by modeling the human brain, and their interpretation. Machine learning algorithms and artificial intelligence applications, which are widely used in many fields, have also begun to be used in real estate valuation. In recent years, statistical-based machine learning regression methods have been a good alternative to other approaches in real estate valuation. Location is undoubtedly the most important factor affecting real estate value. It is necessary to make the data meaningful and interpretable because there are many location factors and the data structure is complex. The use of Geographic Information Systems (GIS) is necessary for spatial analysis when carrying out mass real estate valuation. GIS is a system that enables the collection, visualization, and integration of graphic and attribute information from different data sources, and allows management, analysis, and planning activities to be carried out. By using digital maps and using database and statistical analysis for query purposes, information is classified and helps to interpret the data. GIS provides the opportunity to create databases for mass valuation, facilitate complex analyses, and produce value maps for easy understanding of the valuation process in real estate valuation. A nominal valuation system based on GIS allows users to bring together factors that affect value and use various analysis and query opportunities. In real estate valuation, it is aimed to create an accurate real estate valuation system supported by legal regulations, which eliminates differences in application and brings it to a systematic structure. In this thesis, the aim is to identify the steps to determine the mass and accurate real estate values with a machine learning algorithm-based approach that has an objective perspective and location analysis in real estate valuation. To this end, the study was prepared with the goals of identifying the criteria affecting real estate value, determining the weights of the criteria using machine learning algorithms, producing maps based on location factors affecting real estate value using location analysis, creating a hybrid real estate valuation model using machine learning algorithms and location analysis, and detailing the steps to be followed within the scope of the model to create a methodology. Within the scope of the study, the mass valuation models created using machine learning algorithms in previous studies were examined and their model performance was analyzed. The concepts of real estate valuation were defined and the methods used in mass valuation studies were explained. The development of Property Tax, which is directly related to real estate valuation in the current legislation, was summarized in Türkiye and current problems in Property Tax were evaluated within the framework of legislation. Under the machine learning section, machine learning methods were explained and the most widely used algorithms in machine learning were addressed with their basic working principles. In the model design section, a flowchart for the model created using machine learning algorithms and GIS techniques was created, and the steps determined according to this flowchart were analyzed in detail. Sample location analyses such as proximity, surface, and visibility were carried out in the GIS environment for the model to be used, and the resulting data were combined according to the weights of the factors used in the valuation to produce a value map. Finally, the tools and methods used in the analysis stages using machine learning techniques within the scope of the model were determined.

Benzer Tezler

Tez No
557031
Emlak vergisine esas bir coğrafi bilgi sistemi tasarımı
Design of a geographical information system based on real estate tax
ENİS KALAYCI
Yüksek Lisans
Türkçe
2019
Jeodezi ve Fotogrametri Karadeniz Teknik Üniversitesi
Harita Mühendisliği Ana Bilim Dalı
DOÇ. DR. YAKUP EMRE ÇORUHLU
Tez No
121319
Emlak ve çevre temizlik vergi gelirlerinin KBS ile takibinin yapılmasına yönelik proje tasarımı ve uygulaması
Design and application of an urban information system to monitor real estate and environment cleaning taxes
RUHİ ERMİŞOĞLU
Yüksek Lisans
Türkçe
2002
Jeodezi ve Fotogrametri Gebze Yüksek Teknoloji Enstitüsü
Jeodezi ve Fotogrametri Ana Bilim Dalı
YRD. DOÇ. DR. TAŞKIN KAVZOĞLU
Tez No
601913
Analitik Hiyerarşi Yöntemi (AHP) kullanılarak Coğrafi Bilgi Sistemi (CBS) destekli taşınmaz değer haritası üretimi
Generating Geographical Information System (GIS ) supported real estate value map by using Analytic Hierarchy Process (AHP) method
TANSU ÖZCAN
Yüksek Lisans
Türkçe
2019
Jeodezi ve Fotogrametri Necmettin Erbakan Üniversitesi
Harita Mühendisliği Ana Bilim Dalı
PROF. DR. SÜLEYMAN SAVAŞ DURDURAN
Tez No
534666
CBS tabanlı 3 boyutlu kent modellerinin geliştirilmesi ve bulut bilişim ile sunulması
Developing GIS based 3D city models and sharing with cloud compuing
ŞEVKET BEDİROĞLU
Doktora
Türkçe
2018
Jeodezi ve Fotogrametri Karadeniz Teknik Üniversitesi
Harita Mühendisliği Ana Bilim Dalı
DOÇ. DR. VOLKAN YILDIRIM
Tez No
179845
Taşınmaz değerlemesi bilgi sistemi tasarımı ve uygulaması: Yenişehir örneği
Design of real estate valuation information system and implementation: An example of Yenişehir
FERİHAN ÖZFİDAN
Yüksek Lisans
Türkçe
2008
Jeodezi ve Fotogrametri Zonguldak Karaelmas Üniversitesi
Jeodezi ve Fotogrametri Mühendisliği Bölümü
YRD. DOÇ. DR. MEHMET ALKAN

Geri Dön