KLASTERISASI PROVINSI DI INDONESIA BERDASARKAN PRODUKTIVITAS KOMODITAS PANGAN MENGGUNAKAN ALGORITMA K-MEANS

• Main Author: Aditya Novita, (Author)
• Format: Book
• Published: 2022-07-19.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upnvj.ac.id/19759/1/ABSTRAK.pdf; http://repository.upnvj.ac.id/19759/2/AWAL.pdf; http://repository.upnvj.ac.id/19759/3/BAB%201.pdf; http://repository.upnvj.ac.id/19759/4/BAB%202.pdf; http://repository.upnvj.ac.id/19759/5/BAB%203.pdf; http://repository.upnvj.ac.id/19759/6/BAB%204.pdf; http://repository.upnvj.ac.id/19759/7/BAB%205.pdf; http://repository.upnvj.ac.id/19759/8/DAFTAR%20PUSTAKA.pdf; http://repository.upnvj.ac.id/19759/11/RIWAYAT%20HIDUP.pdf; http://repository.upnvj.ac.id/19759/12/LAMPIRAN.pdf; http://repository.upnvj.ac.id/19759/13/HASIL%20PLAGIARISME.pdf; http://repository.upnvj.ac.id/19759/14/ARTIKEL%20KI.pdf
• Keywords: QA75 Electronic computers. Computer science; QA76 Computer software

This research was conducted with the aim of clustering provinces based on harvested area, production, and productivity of food commodities in Indonesia. Data sourced from the website of the Ministry of Agriculture. The data of this research include data on harvested area, production, and productivity of provinces in Indonesia from 2017 to 2019. The study was conducted using K-Means in grouping a data and evaluated by calculating (Sum of Square Error) SSE in order to find the optimal cluster. This research was executed using Google Collaboratory and the language used was python programming. The results of this provincial clustering study resulted in the optimal cluster at k=3 with a difference in SSE value of 241.05797006047 . The results of clustering in cluster 0 (medium) amount to 29 data with the characteristics that the province has a more dominant variable whose value is lower than cluster 1 and higher than cluster 2, in cluster 1 (high) there are 64 data characteristic of the province having a more dominant variable whose value is higher. higher than clusters 0 and 2, in cluster 2 (low) there are 9 data with provincial characteristics having more dominant variables whose values are lower than clusters 0 and 1.