Marked Log Gaussian Cox Process untuk Menduga Rata-Rata Pengeluaran Per Kapita Per Bulan Rumah Tangga di Kabupaten Kendal Jawa Tengah

Rakhmasari, Aulia Kharis (2023) Marked Log Gaussian Cox Process untuk Menduga Rata-Rata Pengeluaran Per Kapita Per Bulan Rumah Tangga di Kabupaten Kendal Jawa Tengah. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211940000073-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 July 2025. Download (2MB) | Request a copy

Abstract

Pengeluaran per kapita merupakan indikator pengukuran kemiskinan di Indonesia. Setiap 6 bulan sekali, dilaksanakan estimasi pengeluaran per kapita per bulan yang dilakukan oleh Badan Pusat Statistik (BPS) dijalankan melalui Survei Sosial Ekonomi Nasional (SUSENAS) dengan menggunakan metode pendugaan langsung (direct estimation). Pada penelitian ini, ingin dihasilkan estimasi pengeluaran per-kapita hingga level unit terkecil yaitu masing-masing rumah tangga dengan metode Marked Log Gaussian Cox Process. Metode ini diharapkan mampu mendeteksi variasi wilayah yang terdapat di unit yang lebih kecil atau sub-wilayah. Pada penelitan yang dikembangkan sebelumnya melanggar asumsi non-multikolinearitas dan normalitas, oleh karena itu akan dikembangkan model baru yang memenuhi asumsi tersebut. Pada penelitian ini, dengan adanya perbedaan preprocessing data, asumsi multikolinearitas tidak lagi timbul didalam model. Selain itu, model marked yang dikembangkan telah memenuhi asumsi normalitas dan identik, namun masih belum memenuhi asumsi independen. Didapatkan model LGCP terbaik dengan memperhatikan variabel kepadatan penduduk dengan persamaan λ ̂(u)=exp⁡(2.003909+0.4870049.xz (u)) serta model Marked dengan memperhatikan 5 variabel melalui transformasi optimum box-cox dengan nilai λ sebesar -0.26. Melalui model LGCP dan Marked tersebut, terbentuk model Marked LGCP yang memperhatikan pembobotan sampel dan dihasilkan estimasi pengeluaran per-kapita di level terkecil yaitu rumah tangga dan dapat dilihat variasi pengeluaran per-kapita di setiap desa/kelurahan dengan nilai RMSE sebesar Rp1.144.124,00. Terlihat bahwa pengeluaran per-kapita cenderung lebih tinggi di daerah pusat Kabupaten Kendal dan cenderung rendah di daerah selatan dan tenggara Kendal yang merupakan daerah pegunungan.
=======================================================================================================================================
Expenditure per capita is an indicator of poverty measurement in Indonesia. Every 6 months, expenditure estimation is carried out by the Central Bureau of Statistics (BPS) through the National Socioeconomic Survey (SUSENAS) using the direct estimation method. In this study, an estimate of per-capita expenditure to the smallest unit level is to be produced, namely each household with the Marked Log Gaussian Cox Process method. This method is expected to be able to detect regional variations found in smaller units or sub-regions. In the previously developed calculations violated the assumptions of non-multicollinearity and normality, therefore a new model was developed that met those assumptions. In this study, the initial data processing method (pre-processing) was carried out by changing the extraction method of smoothing Nadaraya Watson smoother to equalize data in each village with the same value. Through this method, it was found that there was no multicollinearity in the data. Then, the best LGCP model was obtained by paying attention to the population density variable by λ(u)=exp⁡(2.003909+ 0.4870049.x_z(u)) and the Marked model by paying attention to 5 variables with the response variable undergoing the optimum transformation of Box-C ox with a value λ of -0.26. Marked Model has met normal and identical residual assumptions, but still does not meet independent assumptions. Through the LGCP and Marked models, a Marked LGCP model is formed that considers thinning probability and produces estimates of per-capita expenditure at the smallest level, namely households and can be seen variations in per-capita expenditure in each subdistrict with RMSE equal to Rp1.144.124,00. Per-capita expenditure tends to be higher in the central area of Kendal Regency and tends to be low in the southern and southeastern areas of Kendal which is a mountainous area.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kemiskinan, Marked LGCP, Pengeluaran Per Kapita, Expenditure per capita, Poverty, Small Area Estimation
Subjects:	H Social Sciences > HA Statistics > HA30.6 Spatial analysis H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HA Statistics > HA31.7 Estimation
Divisions:	Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Aulia Kharis Rakhmasari
Date Deposited:	13 Feb 2023 02:28
Last Modified:	13 Feb 2023 02:28
URI:	http://repository.its.ac.id/id/eprint/96899

Actions (login required)

View Item