Konut değerlerini etkileyen kriterlerin regresyon analizi ile incelenmesi
Analysis of the factors that affecting housing values by using regression analysis
- Tez No: 559909
- Danışmanlar: PROF. DR. RAHMİ NURHAN ÇELİK
- Tez Türü: Yüksek Lisans
- Konular: Jeodezi ve Fotogrametri, Şehircilik ve Bölge Planlama, Geodesy and Photogrammetry, Urban and Regional Planning
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2019
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Gayrimenkul Geliştirme Ana Bilim Dalı
- Bilim Dalı: Gayrimenkul Geliştirme Bilim Dalı
- Sayfa Sayısı: 97
Özet
Konut değerlemesinin, ülkemizdeki gayrimenkul değerleme faaliyetleri içinde işlem sayısı olarak büyük payı vardır. Değer tespiti yapılırken, kullanılan verilerin niteliği, bu verilerin kullanılması ve yorumlanması değerleme süreçlerini ve sonuçlarını doğrudan etkilemektedir. Bu sebeple değere etki eden tüm faktörlerin titizlikle incelenmesi önemlidir. Her konutun değeri, karakteristik niteliklerine ve piyasadaki arz talep dengesine göre değişmektedir. Konutun konumuna ilişkin kriterler arasında binanın bulunduğu çevrenin sosyodemografik yapısı, ulaşım imkanları, trafik yoğunluğu, gürültü ve hava kirliliği, şehir merkezine, sosyal kültürel alanlara, yeşil alanlara, iş alanlarına, önemli merkezlere mesafesi sayılabilir. Konuta ait fiziksel özelliklerinden ise alt yapı hizmetlerinden faydalanma durumu, konutun mimari özellikleri, yapı kalitesi, tadilat ihtiyacı, bulunduğu kat, manzarası, cephe sayısı, coğrafi yönü, otoparkı olup olmaması, güvenlik, ısıtma tipi, asansör olup olmaması gibi birçok kriter sıralanabilmektedir. Bu kriterler değerleme çalışmasının amacına göre seçilir ve ulaşılabilen güvenilir verilerle kısıtlıdır. Bu çalışmanın amacı konut değerlerinde farklılığa sebep olan kriterlerin belirlenmesi ve matematiksel olarak ifade edilmesidir. Bir konut sitesi içinde, bağımsız bölümlerin birbirlerine göre üstünlük oluşturan özelliklerini irdeleyebilmek, değer ile aralarındaki ilişkiyi matematiksel olarak ifade edebilmek amacıyla istatistiksel değerleme yöntemlerine başvurulmuştur. Doğrulanmış mevcut piyasa değerlerine dayanarak, değeri özelliklerinin bir fonksiyonu olarak ifade edilmesine imkan veren çoklu regresyon analizi kullanılmıştır. Çoklu regresyon analizi; birden fazla sayıda bağımsız değişkenin kullanılarak, bağımlı değişkenin tahmin edilmesi yöntemidir. Çoklu regresyon analizi SPSS programı aracılığıyla yapılmıştır. Veriler, beş yıldan uzun süredir gayrimenkul değerleme uzmanlığı yapan değerleme uzmanlarının hazırladığı rapordan edinilmiştir. Bütün gayrimenkuller için araştırmalar, yer tespiti, ölçüm ve incelemeleri, piyasa ve çevre araştırmalarının tamamı raporu hazırlayan uzmanlarca yerinde gerçekleştirilmiştir. Tez çalışmasında özellikle manzara kriterlerinin etkileri ve gürültü kirliliğinin yarattığı etkinin incelenmesi amaçlanmış, örneklem seçiminde çalışılması amaçlanan bu kriterlerin mevcut olmasına dikkat edilmiştir. Şehir içinde farklı manzaraların ve gürültü kaynaklarının etkilerini irdeleyebilmek amacıyla bu özelliklere sahip 17 blokta toplam 1.037 konutun yer aldığı bir proje seçilmiştir. Ortak çevresel etkilere sahip olmaları sebebiyle çalışmada ulaşım, önemli merkezlere mesafe gibi özellikler incelenmemiştir. Regresyon analizinde kullanılan değişkenler konumlu olduğu kat, net kullanım alanı, konutun cephe sayısı, sitenin sosyal tesisine yakınlık, coğrafi yön, site girişine ve gürültü kaynaklarına yakınlık olarak belirlenmiştir. Ayrıca manzaraya ilişkin değişkenler, caddeler iki düzey olmak üzere, peyzaj, orman, otopark, duvar ve bu kategoriler dışında kalan diğer karakteristikler olarak yedi ayrı kriter olarak tanımlanmıştır. Değerleme raporunda takdir edilen değerler bağımlı değişken olarak ve belirlenen 14 bağımsız değişken kullanılarak 1.020 konut verisi ile regresyon analizi yapılmıştır. Bu değişkenlerle 3 ayrı model denemesi yapılmış, tüm değişkenlerin dahil edildiği modelin, değeri en yüksek oranda açıkladığı ve tüm değişkenlerin konut değerleri üstünde istatistiksel olarak anlamlı etkiye sahip olduğu görülmüştür. Yapılan regresyon analizi sonucunda, değer üstünde sahip oldukları etkilere göre kriterler; alan, cephe sayısı, kat, peyzaj manzarasına sahip olmak, cadde (2) manzarasına sahip olmak, coğrafi yön, kategorize edilememiş kötü manzaraya sahip olmak, sosyal tesise yakınlık, gürültü, otopark ve trafo manzarasına maruz kalmak, cadde (1) manzarasına sahip olmak, istinat duvarına maruz kalmak, orman manzarasına sahip olmak, site girişine yakın olmak kriterleri olarak sıralanmıştır. Manzara kriterlerinden peyzaj ve orman manzarasına sahip olma değişkenlerinin değer ile pozitif ilişkide olduğu, diğer bir değişle değeri artıran etkiye sahip oldukları, diğer tüm manzara kriterlerinin negatif etkiye sahip olduğu görülmüştür. Site girişine yakın olmanın değere negatif etki yaptığı, sosyal tesise yakın olmanın ise pozitif etki yaptığı yani tercih edilen bir özellik olduğu sonuçları çıkmıştır. Konutun değerinin, bulunduğu kat, cephe sayısı, güneşten faydalanması bakımından coğrafi yön kriterlerinden olumlu etkilendiği, gürültü kaynaklarının ise değer üstünde düşürücü etkisi olduğu sonuçları elde edilmiştir. Oluşturulan regresyon denklemi ile veri seti dışında bırakılan 17 konut için modele dayalı değer tahmini yapılmıştır. Değerleme raporunda takdir edilen değerlere, ortalama %1,8 oranında yakın değerler elde edilmiştir. Kriterler artırılarak farklı özelliklerin tanımlanması ve konuma ilişkin kriterlerin de dahil edilmesi ile daha geniş alanlarda değer tahmini için kullanılabilir modellerin oluşturulabileceği düşünülmektedir.
Özet (Çeviri)
Housing valuation activity plays an important role within the real estate valuation activities in Turkey. In 2018, 867.546 valuation reports were prepared, that includes the valuation of 1.843.253 properties in Turkey. Share of residential real estates and the number of appraised houses in these reports are 50% and 921.582 respectively. Sector reports shows that these numbers and rates were even higher in previous years. As it is understood, valuation of residential properties has the largest share in the real estate appraisal sector, which is a growing sector day by day. The quality of the data used, the interpretation and method of using these data, directly effect the valuation process and its results. Therefore, it is important to examine carefully all the factors effecting the value. The value of each house varies according to the characteristics and the supply and demand equilibrium in the market. The criteria for the location of the house includes the sociodemographic structure of the environment where the building is located, transportation facilities, traffic density, noise and air pollution, distance to the city center, social-cultural areas, green areas, commercial areas, and landmarks. The physical characteristics of the house can be listed in many different criteria such as utilization of infrastructure services, architectural features of the house, quality of the building materials, the need for renovation, the located floor, view, number of facades, geographical direction, parking facilities, security, type of heating, and whether there are elevators or not. Architectural features of the house can be analyzed under the headings such as number of rooms, floor area, living area, kitchen area, number of bathrooms, balcony, car park capacity, garden area, etc.. It is always possible to increase the number of these criteria. The considered criteria depend on the opinions, experience and the research performed by the specialist. It is a general assumption that as the number of criteria taken into consideration increase, the result is becoming more accurate to the real value. However, the criteria used may vary in each valuation study since these criteria should be selected according to the purpose of the valuation study and are limited to accessible reliable data. There are many methods used in real estate appraisal nowadays. The most common methods that practically used are the market approach, the cost approach and the income approach. Besides these most common approaches, there are also some statistical approaches that can be used to express the value of the real estate as a mathematical model. Statistical approaches enable the interpretation of the relationship between value and its factors. It is also possible to examine and mathematically express the valuation differences and thus, allow to differentiate the values of the similar houses which have the same locational and environmental characteristics by using these statistical methods. Multiple regression analysis, hedonic price index and nominal valuation approaches are some of the statistical methods that can be used in order to estimate the analysis of the equation of the factors effecting the values of the real estates. This study includes information about methods and results of previous studies that are related to the aim of this study. The hypothesis and sub-hypotheses of this study were compared with the findings of the previous relevant studies. The aim of this study is to determine the criteria causing differences in housing values and to express them mathematically. Statistical valuation methods were used in order to examine the superior characteristics of different independent residential units that are located in the same residential estate and to express mathematical relationship between their values and characteristics. Based on validated market values, multiple regression analysis was applied in order to enable the expression of value as a function of its properties. Multiple regression analysis; is a method of estimating the dependent variable using more than one independent variables. In this study multiple regression analysis was performed via the SPSS software. The hypothesis of this study is to verify that criteria such as the storey levels, the living area, the number of facades of the house, the aspect, the distance to the social facilities, the distance to the main entrance of the residential unit and the proximity to the noise sources are effective on the house values. There are five sub-hypotheses in this study based on the estimated outcomes of the study. These are; the living area has a positive effect on the price, having a garden or forest view is more preferable than having street and city view, proximity to the noise sources negatively effect the price of the house, the distance to the social facilities and the main entrance positively effect the price of the house. The data used in this study were obtained from the real estate valuation reports prepared by real estate appraisers who have more than five years of experience. All of the due diligence, research, market and environmental surveys, investigations, and measurements in the aforementioned reports were carried out by experts who specialized in this area. Therefore, it shall be considered that reliable accurate information is obtained from real estate valuation reports. One of the specific aims of this thesis is to examine the effects of landscape criteria and noise pollution. In order to examine the effects of different landscapes and noise sources in the city, a project that consists of 1,037 houses with 17 blocks was selected by paying attention to the existence of these criteria. Due to their common environmental impact, the characteristics such as transportation and distance to important city centers could not be examined. The variables used in the regression analysis were determined as the storey levels, the net living area, the number of facades of the house, the aspect, the distance to the social facilities, the distance to the main entrance of the residential unit, the proximity to the noise sources. In addition to this, the variables related to the view were categorized as seven different criteria: streets in two levels, garden, forest, car park, garden wall, and“the undefined”. The undefined is represents a view that cannot be clearly categorized, generally being a bad view. Two valuation reports were prepared, one for all residential units and one for only 482 residential units. 17 houses were excluded from the data set during the analysis for applying the model results later as a test. Thus there are two sample sizes consists of 482 and 1020 houses. Regression analysis was performed with 14 independent variables and market value as the dependent variable. The values of the houses are obtained from the real estate valuation report. 3 different regression models were performed with these variables for two data sets. The model, which all variables were included with the bigger sample, was found to explain the value at the highest level and all variables had a statistically significant effect on the housing values. As a result, this model was considered feasible and applied for excluded houses' value prediction. As a result of regression analysis; the most effective criteria are determined as the net living area, the number of facades, the storey levels, having a garden view, street (number 2) view, the aspect, condition of undefined view respectively. The rest can be sorted as the distance to the social facility, the proximity to the noise sources, car park view, street (number 1) view, the garden wall, the forest view and the distance to the main entrance of the residential unit. It is observed that the garden and forest view variables have a positive relationship with the value, in other words, they have an effect that increases the value. All the other view criteria (street views, garden wall, car park and undefined view) have a negative effect on house value. It has been concluded that being close to the site entrance has a negative effect on the value, while being close to the social facility has a positive effect. Therefore, being close to the social facility is a preferred feature. It is determined that the criteria such as storey level, number of facades, the aspect in term of the utilization of the sun, effect the house values positively. It is also determined that the noise sources have a negative effect that decrease the value of the houses. All these results approve the hypothesis of the study that all the selected criteria are effective on house prices. According to results of the study, there is positive relationship between house price and the living area, having a garden or forest view and distance to the social facilities. Houses that have a garden/forest view are more valuable than the other houses which have a street or city view. The results also show that house price increases as living area increases and house prices begin to rise as the houses get closer to a social facility. It is also determined that the noise sources have a negative effect on the value of the house. Being close to the noise sources has a value decreasing effect which is similar to sub hypothesis about the noise sources. All these results confirm the used sub hypothesis with one exception which is being close to the main enterance that is found to have a positive effect on house value. According to results, there is a negative relationship between the distance to the main enterance and the house price. The reason might be the noise pollution caused by pedestrians and vehicle traffic and commercial units which are located on the enterance. The model-based value estimation was performed for 17 houses that were excluded from the data set by using the regression equation model. The mean difference was found to be average1,8% compared to the predicted values in the model and the report values. The biggest difference is 5,09% and the smallest difference is 0,43% between the predicted values and valuation report results. In this study, it is aimed to determine the criteria that effect the house values, to model them mathematically, to evaluate the relationship between the criteria and the value of the houses, besides the degree of influences of this critera. Where multiple houses are located, the variables that effect the house value are investigated. Therefore, regional impacts such as distance to shopping centers, green spaces, public spaces, transportation and education could not be examined. In addition, common criteria such as security, heating system, garage, age of building, material and workmanship quality, total number of floors in the building and the presence of elevators were not examined within the scope of this study. These are the common features of each house in the study sample. Apart from that, criteria have been determined on the basis of advantages and disadvantages of houses. As a result of the literature research, it was found that view and noise source variables also had significant effects on housing values, therefore the sample was selected where these criterions could be included. In comparision with previous studies, this study contains detailed criteria in a smaller area while it has a larger number of sample. It has been suggested that using more criteria to define different characteristics and inclusion of location criteria in further studies may help to create models that estimate the values in larger areas. Comprehensive models can help to ensure reliable results by enabling cross-check valuation results. Providing open access to valuation experts and all stakeholders in the valuation process can be very useful. Comprehensive models can be used as reference or can be a controlling mechanism during desicion making phase in valuation studies. Also comprehensive models can help making objective evaluations by minimizing the value differences arising from subjective opinions.
Benzer Tezler
- Coğrafi bilgi sistemleri entegreli makine öğrenmesine dayalı toplu taşınmaz değerleme modelinin geliştirilmesi
Development of mass property valuation model based on geographic information systems integrated machine learning methods
MUHAMMED OĞUZHAN METE
Doktora
Türkçe
2022
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiGeomatik Mühendisliği Ana Bilim Dalı
PROF. DR. TAHSİN YOMRALIOĞLU
- Konut değerlemesi ve konut değerlerini etkileyen faktörlerin regresyon analizi ile incelenmesi
Housing valuation and examination of the factors affecting the housing prices by using regression analyse
EBRU YAHŞİ
Yüksek Lisans
Türkçe
2007
Mühendislik Bilimleriİstanbul Teknik ÜniversitesiDisiplinlerarası Ana Bilim Dalı
PROF.DR. VEDİA DÖKMECİ
- CBS tabanlı makine öğrenme teknikleri ile toplu taşınmaz değerlemesi
Mass real estate valuation by GIS-based machine learning techniques
SÜLEYMAN ŞİŞMAN
Yüksek Lisans
Türkçe
2021
Jeodezi ve FotogrametriGebze Teknik ÜniversitesiHarita Mühendisliği Ana Bilim Dalı
PROF. DR. ARİF ÇAĞDAŞ AYDINOĞLU
- Evaluating performance of different remote sensing techniques and various interpolation approaches for soil salinity assessment
Toprak tuzluluğu değerlendirmesi için farklı uzaktan algılama teknikleri ve çeşitli interpolasyon yaklaşımlarının performansının değerlendirilmesi
TAHA GORJI
Doktora
İngilizce
2021
Bilim ve Teknolojiİstanbul Teknik ÜniversitesiBilişim Uygulamaları Ana Bilim Dalı
PROF. DR. AYŞE GÜL TANIK
- Fake news classification using machine learning and deep learning approaches
Makine öğrenimi ve derin öğrenme yaklaşımlarını kullanarak sahte haber sınıflandırması
SAJA ABDULHALEEM MAHMOOD AL-OBAIDI
Yüksek Lisans
İngilizce
2023
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolGazi ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ TUBA ÇAĞLIKANTAR