Space Syntax in Historical Environment with Artificial Intelligence: Typological Classification of Uşak Traditional Houses

Gizem Güvlü Erhan; Seda Şimşek Tolacı

Article Details

Space Syntax in Historical Environment with Artificial Intelligence: Typological Classification of Uşak Traditional Houses

Authors:

Gizem Güvlü Erhan | Seda Şimşek Tolacı

Publisher:

Gazi University

Main Topics:

Bilgi Sistemleri (Diğer) | Mimari Mühendislik

Journal:

Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji Vol.13 Issue 3, pp.1002-1018

Document Type:

Article

Publication Date:

9/2025

Document Language:

Turkish

Document ID (Org No.):

2001353680

pISSN:

2147-9526

Keywords:

Abstract:

 Traditional houses, which are considered as cultural heritage, are not only physical structures but also elements that reflect cultural norms and should be protected in terms of the space-human relations they contain. In order to solve these relations scientifically, the spatial organization logic should be analyzed in detail with contemporary methods. In this study, the spatial organization that is shaped according to cultural differences between communities was examined by integrating the space syntax method, which reveals the numerical identity of the spatial structure and allows us to quantitatively analyze the social logic in the space, with the artificial intelligence algorithms. In order to reveal the spatial reflection of the culture in traditional Turkish houses, it was aimed to combine the classification power of artificial intelligence with the space syntax theory. In the study, morphological analyzes of 12 houses selected from traditional Uşak houses were made and their spatial relations were revealed. The obtained metrics such as depth, integration, and connectivity were processed in the Python environment and an artificial intelligence-supported classification model was created. Typology estimates were made using the XGBoost algorithm based on a supervised learning approach. The findings show that the spatial structure of traditional houses can be successfully separated by artificial learning systems. This method offers a new methodological approach to the digital analysis of cultural heritage.

 Traditional houses, which are considered as cultural heritage, are not only physical structures but also elements that reflect cultural norms and should be protected in terms of the space-human relations they contain. In order to solve these relations scientifically, the spatial organization logic should be analyzed in detail with contemporary methods. In this study, the spatial organization that is shaped according to cultural differences between communities was examined by integrating the space syntax method, which reveals the numerical identity of the spatial structure and allows us to quantitatively analyze the social logic in the space, with the artificial intelligence algorithms. In order to reveal the spatial reflection of the culture in traditional Turkish houses, it was aimed to combine the classification power of artificial intelligence with the space syntax theory. In the study, morphological analyzes of 12 houses selected from traditional Uşak houses were made and their spatial relations were revealed. The obtained metrics such as depth, integration, and connectivity were processed in the Python environment and an artificial intelligence-supported classification model was created. Typology estimates were made using the XGBoost algorithm based on a supervised learning approach. The findings show that the spatial structure of traditional houses can be successfully separated by artificial learning systems. This method offers a new methodological approach to the digital analysis of cultural heritage.

 Traditional houses, which are considered as cultural heritage, are not only physical structures but also elements that reflect cultural norms and should be protected in terms of the space-human relations they contain. In order to solve these relations scientifically, the spatial organization logic should be analyzed in detail with contemporary methods. In this study, the spatial organization that is shaped according to cultural differences between communities was examined by integrating the space syntax method, which reveals the numerical identity of the spatial structure and allows us to quantitatively analyze the social logic in the space, with the artificial intelligence algorithms. In order to reveal the spatial reflection of the culture in traditional Turkish houses, it was aimed to combine the classification power of artificial intelligence with the space syntax theory. In the study, morphological analyzes of 12 houses selected from traditional Uşak houses were made and their spatial relations were revealed. The obtained metrics such as depth, integration, and connectivity were processed in the Python environment and an artificial intelligence-supported classification model was created. Typology estimates were made using the XGBoost algorithm based on a supervised learning approach. The findings show that the spatial structure of traditional houses can be successfully separated by artificial learning systems. This method offers a new methodological approach to the digital analysis of cultural heritage.

 Kültürel miras niteliği taşıyan geleneksel konutlar, yalnızca fiziksel yapıları değil barındırdıkları mekan-insan ilişkileri bakımından korunması gereken, kültürel normların yansıtıcısı görevini üstlenen ögelerdir. Bu ilişkilerin bilimsel biçimde çözümlenebilmesi için mekânsal örgütlenme mantığının çağdaş yöntemler ile detaylı bir şekilde analiz edilmesi gerekmektedir. Bu çalışmada mekansal kurgunun sayısal kimliğini ortaya çıkaran ve mekandaki sosyal mantığı nicel olarak analiz etmemizi sağlayan mekan dizimi yöntemi ile topluluklar arasındaki kültürel farklılıklara göre şekillenen mekânsal kurgular yapay zeka algoritmaları ile bütünleştirilerek incelenmiştir. Geleneksel Türk konutlarındaki kültürün mekânsal yansımasının ortaya koyulabilmesi için mekan dizim teorisi ile yapay zekanın sınıflandırma gücünü birleştirmek amaçlanmıştır. Çalışmada geleneksel Uşak evlerinden seçilen 12 konutun morfolojik analizleri yapılarak mekânsal ilişkileri ortaya konulmuştur. Elde edilen derinlik, entegrasyon, bağlanırlık gibi metrikler Python diliyle işlenerek yapay zekâ destekli sınıflandırma modeli oluşturulmuştur. Denetimli öğrenme yaklaşımına dayalı XGBoost algoritması kullanılarak tipoloji tahminleri gerçekleştirilmiştir. Bulgular, geleneksel konutların mekânsal yapısının yapay öğrenme sistemleri tarafından başarılı şekilde ayrıştırılabildiğini göstermektedir. Bu yöntem, kültürel mirasın dijital analizine yönelik yeni bir metodolojik yaklaşım sunmaktadır.

Article Details

Space Syntax in Historical Environment with Artificial Intelligence: Typological Classification of Uşak Traditional Houses

Authors:

Gizem Güvlü Erhan | Seda Şimşek Tolacı

Publisher:

Gazi University

Main Topics:

Bilgi Sistemleri (Diğer) | Mimari Mühendislik

Journal:

Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji Vol.13 Issue 3, pp.1002-1018

Document Type:

Article

Publication Date:

9/2025

Document Language:

Turkish

Document ID (Org No.):

2001353680

pISSN:

2147-9526

Space Syntax

Artificial Intelligence

Machine Learning

Uşak Traditional HousesMekan Dizim

Yapay Zeka

Makine Öğrenmesi

Uşak Geleneksel Konutları

Issues

Article Details

Authors: Gizem Güvlü Erhan | Seda Şimşek Tolacı

Publisher: Gazi University

Main Topics: Bilgi Sistemleri (Diğer) | Mimari Mühendislik

Journal: Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji Vol.13 Issue 3, pp.1002-1018

Document Type: Article

Publication Date: 9/2025

Document Language: Turkish

Document ID (Org No.): 2001353680

pISSN: 2147-9526

Issues

Authors:

Gizem Güvlü Erhan | Seda Şimşek Tolacı

Publisher:

Gazi University

Main Topics:

Bilgi Sistemleri (Diğer) | Mimari Mühendislik

Journal:

Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji Vol.13 Issue 3, pp.1002-1018

Document Type:

Article

Publication Date:

9/2025

Document Language:

Turkish

Document ID (Org No.):

2001353680

pISSN:

2147-9526