DETECTING URBAN SLUMS IN DKI JAKARTA: A KOTAKU DATA APPROACH WITH ENSEMBLE METHODS
Slums are one of the problems that often occur in urban areas, especially in developing countries. Slum settlements cause various social, economic, and environmental problems, including social injustice, infrastructure inefficiency, and a decrease in the population's quality of life. The PUPR M...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universitas Pattimura
2024-07-01
|
| Series: | Barekeng |
| Subjects: | |
| Online Access: | https://ojs3.unpatti.ac.id/index.php/barekeng/article/view/12075 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Slums are one of the problems that often occur in urban areas, especially in developing countries. Slum settlements cause various social, economic, and environmental problems, including social injustice, infrastructure inefficiency, and a decrease in the population's quality of life. The PUPR Ministry representing the Indonesian government is trying to overcome slum settlements in Indonesia by creating the Cities Without Slums (KOTAKU) program. The KOTAKU program provides relevant and detailed data on slum settlements in Indonesia. Challenges arise when analyzing and utilizing KOTAKU data to identify slum indicators and map slums broadly. The method used in detecting slums using KOTAKU data is still conventional. Machine learning can be used to model data and classify or predict data by applying the Ensemble Method. This modeling will look for patterns or structures from the data that has been provided so that the detection results become more objective. This study aims to model slum indicators from KOTAKU data and detect urban slum settlements in DKI Jakarta. Modeling is done using the Random Forest algorithm. Data sourced from the KOTAKU program website established by the Ministry of PUPR RI. The results of the study show that the indicators that contribute most to the modeling of urban slum indicators in DKI Jakarta are the availability of safe access to drinking water and not fulfilling needs for drinking water. The slum indicator model without additions has good performance after going through the parameter tuning process with parameters ntree = 500 and mtry = 6. In contrast, the slum indicator model with additions has good performance if it does not go through a parameter tuning process or retains its initial parameters namely ntree = 500 and mtry = 4. |
|---|---|
| ISSN: | 1978-7227 2615-3017 |