Electronic Theses and Dissertation

Zona Nilai Tanah (ZNT) merupakan dasar dalam penentuan Nilai Jual Objek Pajak (NJOP) yang digunakan sebagai acuan dalam penilaian tanah. Pembaruan ZNT umumnya dilakukan melalui analisis spasial secara manual sehingga berpotensi menghasilkan peta ZNT yang belum sepenuhnya mencerminkan dinamika nilai pasar aktual. Penelitian ini bertujuan mengklasifikasikan ZNT di Kecamatan Syiah Kuala, Kota Banda Aceh dengan mengintegrasikan Sistem Informasi Geografis (SIG) dan metode Support Vector Machine (SVM) Multiclass. Data penelitian terdiri atas 12.492 persil tanah. Pembentukan variabel dilakukan melalui analisis proximity menggunakan Near Tool pada ArcGIS untuk menghitung jarak setiap persil terhadap 13 objek spasial yang memengaruhi nilai tanah. Selanjutnya, data dibagi menggunakan stratified sampling dengan proporsi 80% data pelatihan dan 20% data pengujian, kemudian distandarisasi menggunakan StandardScaler. Model SVM Multiclass dibangun menggunakan parameter terbaik yang diperoleh melalui Grid Search Cross Validation dan dievaluasi menggunakan confusion matrix, accuracy, precision, recall, dan F1-score. Hasil penelitian menunjukkan bahwa model memperoleh nilai accuracy sebesar 96,64%, precision sebesar 96,69%, recall sebesar 96,64%, dan F1-score sebesar 96,64%. Hasil klasifikasi juga menunjukkan pola persebaran yang sangat mendekati peta ZNT eksisting dengan hanya 84 persil yang mengalami kesalahan klasifikasi dari 2.499 data pengujian. Penelitian ini menunjukkan bahwa integrasi analisis proximity menggunakan Near Tool dan metode SVM Multiclass mampu menghasilkan klasifikasi ZNT yang akurat serta berpotensi mendukung proses pembaruan ZNT secara lebih efisien.

Kata kunci: Analisis Proximity, Near Tool, Sistem Informasi Geografis, Support Vector Machine Multiclass, Zona Nilai Tanah.

Land Value Zones (LVZ) serve as the basis for determining the Assessed Value of Tax Objects (NJOP), which is used as a reference for land valuation. Updating Land Value Zones is generally conducted through manual spatial analysis, which may result in LVZ maps that do not fully reflect the dynamics of current market values. This study aims to classify Land Value Zones in Syiah Kuala District, Banda Aceh City, by integrating Geographic Information Systems (GIS) with the Multiclass Support Vector Machine (SVM) method. The study employed 12,492 land parcels as the research dataset. The research variables were generated through proximity analysis using the Near Tool in ArcGIS to calculate the distance from each land parcel to 13 spatial features that influence land values. Subsequently, the dataset was divided into training and testing sets using stratified sampling with a ratio of 80:20 and standardized using StandardScaler. The Multiclass SVM model was developed using the optimal hyperparameters obtained through Grid Search Cross Validation and evaluated using a confusion matrix, accuracy, precision, recall, and F1-score. The results showed that the proposed model achieved an accuracy of 96.64%, precision of 96.69%, recall of 96.64%, and F1-score of 96.64%. The classification results also demonstrated a spatial distribution that closely resembled the existing Land Value Zone map, with only 84 misclassified land parcels out of 2,499 testing samples. These findings indicate that the integration of proximity analysis using the Near Tool and the Multiclass Support Vector Machine method is capable of producing accurate Land Value Zone classifications and has the potential to support a more efficient Land Value Zone updating process. Keywords: Geographic Information System, Land Value Zone, Multiclass Support Vector Machine, Near Tool, Proximity Analysis.