Hidayati, Rosikhu Ilmi (2019) Analisis Imbalanced Multiclass Pada Status Kepemilikan Asuransi Dengan Metode Multinomial Logistic Regression. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211540000077_Undergraduate_Theses.pdf Download (3MB) | Preview |
Abstract
Masalah kesehatan di Indonesia sangat bervariasi dan untuk mengatasinya pemerintah menggalakkan program asuransi. Saat ini asuransi berkembang pesat akan tetapi perkembangan itu tidak diiringi oleh pemahaman masyarakat akan pentingnya asuransi. Hal itu tampak masih minimnya jumlah nasabah yang terdaftar sebagai peserta asuransi. Tidak seperti negara maju di barat dimana asuransi menjadi kewajiban bagi seluruh warganya. Mengingat asuransi cukup penting maka dilakukan analisis status kepemilikan asuransi dengan metode regresi logistik multinomial. Untuk mengatasi kasus imbalanced data dimana status kepemilikan asuransi cenderung pada satu kategori, maka dilakukan balancing data dengan SMOTE, random undersampling dan combine sampling. Hasil analisis menunjukkan bahwa analisis dengan balancing undersampling lebih baik dibandingkan metode yang lain karena model yang dihasilkan fit dengan nilai AUC sebesar 54,78%, dimana status kepemilikan asuransi dipengaruhi oleh lokasi tempat tinggal dan frekuensi rawat jalan. Diharapkan hasil analisis ini dapat memberikan informasi bagi pemerintah dalam upaya meningkatkan SDM dibidang kesehatan. ================================================================================================================================
Health problems in Indonesia are very diverse and to overcome them the government promotes insurance programs. Currently, insurance is growing rapidly but the development is not accompanied by a public understanding of the importance of insurance. It seems that there is still a minimum number of customers registered as insurance participants. Unlike developed countries in the west where insurance is an obligation for all citizens. Given that insurance is important, an analysis of insurance ownership status is carried out with multinomial logistic regression method. To overcome imbalanced data where insurance ownership status tends to be in one category, preprocessing is done with SMOTE, undersampling and combine sampling. The analysis shows that balancing data with undersampling is better than the other methods because the model is fitted with AUC 54,78%, where the ownership status of insurance is influenced by the location of residence and frequency of outpatient care. It is expected that the results of this analysis can provide information for the government in an effort to improve human resources in the health sector.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.536 Hid a-1 2019 |
| Uncontrolled Keywords: | Asuransi, AUC, Imbalanced Data, Regresi Logistik Multinomial. |
| Subjects: | H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HG Finance > HG4012 Mathematical models H Social Sciences > HG Finance > HG8771 Life insurance Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Rosikhu Ilmi Hidayati |
| Date Deposited: | 16 Jun 2023 02:53 |
| Last Modified: | 16 Jun 2023 02:53 |
| URI: | http://repository.its.ac.id/id/eprint/64492 |
Actions (login required)
 |
View Item |