Abstract:

The unplanned growth of cities has brought about unplanned urbanization over time. To generate a regular city model, municipalities need to transform the areas with irregular cadastral parcels into standard parcels, which is possible with zoning regulations. One of the methods used to generate social life areas that arise when people start to become larger communities in urbanization is the application of Article 18 of the Zoning Law No. 3194. The municipalities make this application ex officio to switch to more regular urbanization. The zoning parcels the cadastral parcel will be given in the distribution process is done manually. This situation prolongs the implementation period and causes the number of partners to be far from the optimum in the newly assigned parcels. A solution has been sought for the distribution problem, clustering problem characteristically, which is considered within the scope of the study, with K-means and Adaptive K-means Algorithms, which are clustering method within the scope of spatial data mining. Twelve cadastral parcel areas were distributed to 6 zoning parcels. As a result of the experimental evaluations, it was seen that the Adaptive K-means algorithm was more successful. As a result of the experimental assessment, while the minimum area that cannot be distributed in the parcels was 435 km2 in the Classical K-means method, this ratio decreased to 23 km2 in the Adaptive K-means method, and the Adaptive K-means method came to the fore in the general evaluation.