Abstract
The number of traffic accidents in Bali kept increasing since 2015 until 2017. The factors that affected the traffic accidents in every region were suspected to be varied according to geographic position. This geographic effect was known as spatial heterogeneity. Spatial heterogeneity was analized by using Geographically Weighted Regression (GWR). This study aim to model the factors which affected the traffic accidents in every subdistrict in Bali by using fixed and adaptive gaussian kernel. The result showed that GWR with adaptive gaussian kernel was better at estimated the models because it had higher value of which was at . The factors which significantly affected the number of traffic accident in 57 subdistrict in Bali were the average rainfall and the number of population within age of 15 to 29 years old.
Highlights
This geographic effect was known as spatial heterogeneity
The result showed that Geographically Weighted Regression (GWR) with adaptive gaussian kernel was better at estimated the models because it had higher value of which was at
The factors which significantly affected the number of traffic accident in 57 subdistrict in Bali were the average rainfall and the number of population within age of 15 to 29 years old
Summary
Bali merupakan salah satu dari 34 Provinsi yang ada di Indonesia dengan jumlah penduduk 4.246.500 jiwa dan luas wilayah 5.636,66 km, sehingga kepadatan penduduk di Bali mencapai 753 jiwa/km (BPS, 2018). Melihat kasus kecelakaan di Bali terus mengalami peningkatan dan perbedaan karakteristik yang dimiliki oleh masing-masing kecamatan, maka dalam penelitian ini diharapkan diketahui hubungan kecelakaan lalu lintas dan faktor-faktor yang memengaruhinya di setiap kecamatan yang ada di Provinsi Bali. Pada umumnya untuk mengetahui model hubungan antara variabel bebas dengan variabel terikat digunakan metode regresi linier (Dewi & Zain 2016). Variabel bebas yang digunakan adalah kepadatan penduduk, persentase usia remaja, persentase kecelakaan terjadi di kawasan jalan kabupaten/kota, persentase pendidikan terakhir pelaku adalah di atas SMP, rasio jenis kelamin dan persentase kecelakaan terjadi pada waktu gelap. Hasil uji kesesuaian model menyatakan bahwa tidak ada perbedaan antara model regresi linier berganda dengan model GWR namun model GWR dengan fungsi pembobot fixed kernel gaussian lebih sesuai digunakan. Berdasarkan paparan di atas, penulis melakukan penelitian mengenai faktor-faktor yang memengaruhi kecelakaan lalu lintas pada masing-masing kecamatan yang ada di Provinsi Bali menggunakan metode GWR
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call