Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/165677| Title: | Penerapan Algoritme K-Prototypes untuk Segmentasi Mahasiswa Baru Program Sarjana IPB University Tahun 2024 Berdasarkan Karakteristiknya |
| Other Titles: | Implementation of the K-Prototypes Algorithm for Segmenting New Undergraduate Students at IPB University in 2024 Based on Characteristics |
| Authors: | Soleh, Agus Mohamad Dito, Gerry Alfa PRATIWI, OKTAVIA GALIH |
| Issue Date: | 2025 |
| Publisher: | IPB University |
| Abstract: | IPB University setiap tahun menerima mahasiswa baru dari latar belakang beragam sehingga muncul tantangan dalam penyediaan layanan yang adil dan tepat sasaran. Salah satu pendekatan untuk mengatasi tantangan tersebut adalah segmentasi mahasiswa menggunakan algoritme K-Prototypes, yang dirancang untuk menangani data campuran dengan menggabungkan prinsip K-Means (numerik) dan K-Modes (kategorik). Penelitian ini menerapkan algoritme K-Prototypes untuk segmentasi mahasiswa baru program sarjana IPB University tahun 2024 dan mengidentifikasi karakteristik mahasiswa dari hasil penggerombolan yang optimal. Data mencakup 4723 mahasiswa dengan empat peubah numerik dan sebelas peubah kategorik. Proses penggerombolan dilakukan setelah praproses data, standarisasi, penyederhanaan kategori peubah, dan dievaluasi menggunakan rasio S_W/S_B minimum. Hasil penelitian menunjukkan jumlah gerombol optimal adalah lima dengan rasio S_W/S_B sebesar 0,121. Profil gerombol mencerminkan keragaman kondisi sosial ekonomi mahasiswa mulai dari kelompok rentan hingga kelompok paling mapan. Segmentasi ini diharapkan menjadi masukan strategis dalam penetapan kebijakan UKT, bantuan keuangan, serta strategi penerimaan mahasiswa baru di IPB University. IPB University annually admits new undergraduate students from diverse backgrounds, posing challenges in delivering fair and targeted services. One approach to overcoming this issue is student clustering using the K-Prototypes algorithm, which is designed to handle mixed-type data by combining the principles of K-Means (for numerical data) and K-Modes (for categorical data). This study applies the K-Prototypes algorithm to segment new undergraduate students at IPB University in 2024 and identifies the characteristics of the optimal clustering results. The dataset comprises 4,723 students with four numerical variables and eleven categorical variables. Clustering was performed following data preprocessing, variable recoding, and standardization, with evaluation using the minimum S_W/S_B ratio. The study results indicate that the optimal number of clusters is five, indicated by the minimum S_W/S_B ratio of 0,121. The cluster profiles reflect the diversity of socio-economic conditions, ranging from vulnerable groups to the most affluent. This segmentation is expected to serve as strategic input for setting tuition fee (UKT) policies, financial aid programs, and new student admission strategies at IPB University. |
| URI: | http://repository.ipb.ac.id/handle/123456789/165677 |
| Appears in Collections: | UT - Statistics and Data Sciences |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| cover_G1401211066_81548132c873448fb8f08e311cdd7d07.pdf | Cover | 587.13 kB | Adobe PDF | View/Open |
| fulltext_G1401211066_17ca7d172e1c45a29fab17c5b89d8ea3.pdf Restricted Access | Fulltext | 1.51 MB | Adobe PDF | View/Open |
| lampiran_G1401211066_7231bcf489b64a55b9cfe89a0a74421e.pdf Restricted Access | Lampiran | 483.37 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.