In this study, a new methodology has been developed for a sustainable mass appraisal system. A mathematical model was created with the combination of the Cobb-Douglas and the linear regression model. With the Analytic Hierarchy Process (AHP) method, real estate value criteria were grouped and weighted in a hierarchical structure. The weights obtained with AHP were integrated into the coefficients regarding the criteria weights and densities in the Cobb-Douglas hybrid model. The new hybrid model was confirmed with the features and price equivalents of 435 parcels for sale from the market. Besides, the model analysis results were compared with the Multiple Regression Analysis (MRA) modelling using market prices. While creating the methodology, Geographic Information Systems (GIS) was used to organize the geographic and regional data of the region. After developing the new hybrid model, criteria groups that developed the model and relevant sub-criteria were evaluated using Pearson's correlation analysis.

In real estate, mass appraisal is a very important subject in the valuation of two and more properties. It can be of benefit in a number of fields including taxation, banking transactions, expropriation, etc. The base problem is which criteria to use for mass appraisal. Because the number of criteria and the criteria themselves vary according to people, regions and countries, they are uncertain. They should be optimum in order to save on time, labor and cost. The aim of this study is to reduce the criteria by determining which ones affect the plot value. A survey which was answered by a total of 2,531 participants was conducted in Turkey. Principal Component Analysis (PCA), one of the criteria analysis methods, was applied to the survey data. The number of criteria was reduced to 14 with separation and to 30 according to the results of PCA. But they decreased in the model verification when criteria data for 558 samples were collected in the Konya study area. An index of the neighborhood and locational features of these criteria was created by using GIS. Three models were established using Multiple Regression Analysis (MRA) and the performance of the models was examined. The prediction values and the market value were integrated into the GIS to compare the spatial distributions of plot values.

Considering water quality is an essential requirement in terms of environmental planning and management. To protect and manage water resources effectively, it is necessary to develop an analytical decision-support system. In this study, a systematic approach was suggested to evaluate the lake water quality. The methodology includes the prediction of the values in different locations of the lakes from experimental data through inverse distance weighting (IDW) method, creation of maps by using Geographic Information System (GIS) integrated with analytic hierarchy process (AHP) from multi-criteria decision analysis (MCDA), reclassification into five class, combining the time-related spatial data into a single map to predict the whole lake water quality from the data of sampling points, and finally overlapping the final maps with topography/geology and land use. The proposed approach was verified and presented as case study for Meke and Acigol Lakes in Konya/Turkey which were affected by human and natural factors although they have ecological, hydromorphological, and socio-economic importance. In the proposed approach, categorizing water quality parameters as “hardness and minerals,” “substrates and nutrients,” “solids content,” “metals,” and “oil-grease” groups was helpful for AHP with the determined group weights of 0.484, 0.310, 0.029, and 0.046, respectively. Assigning weights within each group and then assigning weights between groups resulted in creating accurate final map. The proposed approach is flexible and applicable to any lake water quality data; even with a limited number of data, the whole lake water quality maps could be created for assessment.

The use of machine learning in real estate is quite new. When the working area is large, the factors affecting the price may vary according to the geographical regions and socioeconomic factors. It is thought that the price prediction performance of a model that will reflect these differences will be more successful than a general model. Unsupervised learning methods can be used both to increase performance and to show the variation of different factors affecting the price according to regions. With this aim, a hybrid model of X-Means clustering and CART decision trees was established in this study. This model successfully learned the geographical and physical variables that affect the price. The prediction performance of the model was compared with the direct capitalization method, which is the gold standard in the domain. The hybrid model has a superior performance over direct capitalization in terms of mean square error, root mean square error and adjusted R-Squared metrics. The scores were 72.86, 0.0057 and 0.978, respectively. The effect of clustering was also examined. Clustering increased the prediction performance by 36%.

Real estate is a form of immovable asset that enables individuals to exert their property rights and provides a form of material guarantee through its economic value. The economic value of real estate, which is reflected in the price, is an aspect that all countries emphasize today by identifying purchase and sale values in market conditions that are removed from large project rumors and speculations. The more the real estate market value is removed from reality, the more negative its effect will be on the cost-benefit of real estate management. This study aimed to identify the main criteria that affect parcel value, which constitutes a basis for real estate, narrow them down to the optimum level using questionnaires and standardize them. A total of 559 experts working in real estate valuation and 1,915 members of the public that play a role in real estate purchases and sales were contacted in Ankara, Konya and Kayseri, all of which are located in the Central Anatolia Region of Turkey. The factor analysis method was applied to the survey data. Grouping was carried out with 10 factors and the results were interpreted.