Penerapan algoritma Expectation-Maximization (EM) untuk mengestimasi parameter probabilitas pada kasus dua koin dengan data yang tidak diketahui asalnya. Melalui proses iteratif, diperoleh nilai parameter yang konvergen dan digunakan untuk menentukan kecenderungan koin pada setiap set percobaan.

Laporan ini disusun untuk memenuhi tugas mata kuliah Topik dalam Statistika I

Bozza bollettino n. 1 - 2026

This case study examines the factors affecting procurement delivery performance at Petrohawk Centrum Limited, an oil and gas servicing company. The business problem is the occurrence of late procurement deliveries, which can delay project execution, increase costs, and reduce client satisfaction. The dataset collected contains 100 purchase order observations, including purchase order date, product/item, item category, vendor, supplier source, quantity, unit price, revenue/order value, planned delivery days, and delivery status. The analysis applies the five required Case Study 1 techniques: exploratory data analysis, visualisation, hypothesis testing, correlation analysis, and regression, as specified in the assessment brief. Initial review of the data shows that 72% of deliveries were on time while 28% were late. Late deliveries were more common among international suppliers, higher-value orders, larger quantities, and categories such as packaging, textile/linen, and medical equipment. The study recommends a data-driven supplier performance review system, tighter monitoring of high-risk item categories, and improved lead-time planning for international and high-value procurement orders.

Simulasi Monte Carlo digunakan untuk memprediksi permintaan dengan memanfaatkan bilangan acak dan proses simulasi berulang. Percobaan dilakukan dengan berbagai periode dan jumlah iterasi untuk melihat pengaruhnya terhadap kestabilan dan hasil prediksi.