Prediksi Konsumsi Rumah Tangga Di Kota Surabaya Menggunakan Metode Random Forrest Regression

Basuni, Mohammad Bagussurya (2023) Prediksi Konsumsi Rumah Tangga Di Kota Surabaya Menggunakan Metode Random Forrest Regression. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211940000078-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy

Abstract

Kenaikan penduduk miskin di Kota surabaya antara tahun 2019 hingga 2021 sebanyak 21,940. Hal ini wajib menjadi perhatian pemerintah Kota Surabaya. Dalam upaya pengentasan kemiskinan dibutuhkan informasi jumlah penduduk yang dikategorikan miskin secara akurat. Pada proses pengumpulan data, sering terjadi bahwa tiap rumah tangga tidak memberikan secara detail atau benar. Hal tersebut dapat diantisipasi menggunakan metode PMT (Proxy Mean Test). PMT digunakan untuk memperkirakan kondisi sosial ekonomi setiap rumah tangga (ruta) menggunakan data karakteristik rumah tangga yang mudah diukur dan sulit untuk dimanipulasi. Meninjau penelitian terdahulu kedua penelitian didapatkan hasil yang tidak memenuhi asumsi klasik regresi sehingga model kurang dapat diandalkan. Oleh karena itu, dalam penelitian ini saya menggunakan pendekatan machine learning yaitu random forest regression. Setelah ditemukan hasil prediksi dari metode random forest regression, dapat terlihat pada scatter plot yang dihasilkan terdapat 2 titik dengan jarak yang jauh dibandingkan titik-titik lainnya. Hal ini mengindikasikan adanya outlier pada data sehingga dilakukan penghapusan outlier dengan z sama dengan 3. Setelah dilakukan penghapusan outlier, pada metode random forest regression terjadi penurunan performa dilihat dari nilai RMSE nya yang meningkat sedikit. Ditentukan model terbaik dengan membandingkan nilai RMSE. Ditemukan bahwa hasil prediksi terbaik dihasilkan oleh metode random forest regression dengan adanya outlier dengan nilai RMSE 829.480,84.
=================================================================================================================================
There is an increase in the number of poor people in Surabaya City between 2019 and 2021 by 21,940. This is one of the Surabaya City government concerns. In poverty alleviation efforts, accurate information on the number of people categorized as poor is needed. In the data collection process, it often happens that each household does not provide details of their expenditure correctly. This can be anticipated using the PMT (Proxy Mean Test) method. PMT is used to estimate the socio-economic condition of each household using data on household characteristics that are easy to measure and difficult to manipulate. Reviewing previous studies, the two studies obtained results that did not meet the classical assumptions of regression, so the model was less reliable. Therefore, in this study I used a machine learning approach, namely random forest regression. After finding the best prediction from the random forest regression, it can be seen in the resulting scatter plot that there are 2 points that are farther apart than the other points. This indicates that there are outliers in the data so that the outliers are removed with z equal to 3. After removing the outliers, the random forest regression method shows a decrease in performance as seen from the RMSE value which increases slightly. The best model is determined by comparing the RMSE values. It was found that the best predictive results were produced by the random forest regression method with an outlier. This model has RMSE value of 829,480.84.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Consumption, PMT, Poor, Random Forest Regression, RMSE, Konsumsi, Miskin
Subjects:	H Social Sciences > HA Statistics > HA31.3 Regression. Correlation Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Mohammad Bagussurya Basuni
Date Deposited:	15 Sep 2023 07:01
Last Modified:	15 Sep 2023 07:01
URI:	http://repository.its.ac.id/id/eprint/104597

Actions (login required)

View Item