Prediksi Nilai Tanah Menggunakan Support Vector Regression (SVR) di Kota Surabaya

Fajarianto, Muhammad Rayhan Bayu (2026) Prediksi Nilai Tanah Menggunakan Support Vector Regression (SVR) di Kota Surabaya. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5016221033-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Nilai tanah merupakan komponen penting dalam administrasi pertanahan dan perencanaan wilayah, namun pendekatan konvensional berbasis Zona Nilai Tanah (ZNT) cenderung bersifat generalisasi wilayah sehingga kurang mampu merepresentasikan variasi nilai tanah secara spasial. Penelitian ini bertujuan mengembangkan model prediksi Nilai Bidang Tanah (NBT) di wilayah kerja Kantor Pertanahan Surabaya II menggunakan algoritma Support Vector Regression (SVR) berbasis variabel spasial sebagai alternatif pendekatan tersebut. Model dibangun menggunakan 833 data transaksi bidang tanah tahun 2023-2025 dengan variabel input berupa jarak terhadap jaringan jalan, fasilitas pendidikan, kesehatan, perdagangan, CBD, dan POI, kepadatan penduduk, serta zona RDTR, kemudian dioptimasi menggunakan Bayesian Optimization dan dievaluasi menggunakan K-Fold Cross Validation. Hasil penelitian menunjukkan model mampu menjelaskan variasi nilai tanah dengan R² sebesar 0,53, dengan zona RDTR sebagai variabel paling berpengaruh terhadap prediksi, serta nilai tanah tertinggi terkonsentrasi di kawasan Sukolilo, Simokerto, dan Mulyorejo.
===============================================================================================================================================
Land value is a critical component in land administration and regional planning, yet the conventional approach based on Land Value Zones (ZNT) tends to generalize spatial units, limiting its ability to represent land value variation at the parcel level. This study aims to develop a land parcel value prediction model for the working area of the Surabaya II Land Office using a Support Vector Regression (SVR) algorithm based on spatial variables as an alternative approach. The model was built using 833 land transaction records from 2023–2025, with input variables including distances to road networks, educational, health, and commercial facilities, CBD, and POI, population density, and RDTR zoning, optimized using Bayesian Optimization and evaluated using K-Fold Cross Validation. The results show that the model explains land value variation with an R² of 0.53, with RDTR zoning identified as the most influential variable in the prediction, and the highest land values concentrated in the Sukolilo, Simokerto, and Mulyorejo areas.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Nilai Tanah, Support Vector Regression, Variabel Spasial, Prediksi. Land Value, Support Vector Regression, Spatial Variables, Prediction.
Subjects:	G Geography. Anthropology. Recreation > GA Mathematical geography. Cartography > GA109.5 Multipurpose cadastres.
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geomatics Engineering > 29202-(S1) Undergraduate Thesis
Depositing User:	Muhammad Rayhan Bayu Fajarianto
Date Deposited:	24 Jul 2026 07:01
Last Modified:	24 Jul 2026 07:01
URI:	http://repository.its.ac.id/id/eprint/137991

Actions (login required)

View Item