SEMIVARIOGRAM MODELLING ON MULTIVARIATE DATA USING PRINCIPAL COMPONENT ANALYSIS (CASE STUDY: GOLD PROSPECT DATA IN ACEH)

• Main Author: Chandra, Benny
• Format: Final Project
• Language: Indonesia
• Subjects: Matematika
• Online Access: https://digilib.itb.ac.id/gdl/view/33760
• Institution: Institut Teknologi Bandung

In a mining, it is not uncommon to search an mineral element that has a high economical value such as gold to earn huge profits. On the gold prospects data in Aceh, did the gold prospecting but found also elements of minerals such as silver, copper, and other elements that have a spatial correlation with gold. It would take a very long time and much energy if we want to know the spatial correlation on each element in the data. From the data obtained information that gold has a fairly strong correlation with silver and bismuth. To reduce the amount of variables into fewer without losing the variability of the data, do a technique called principal component analysis. Retrieved nine principal components which absorb more than half of the diversity of the original data. Spatial correlation between locations on variables obtained from the principal component analysis can be described by semivariogram. The experimental semivariogram is calculated based on yield variables outcome of principal component analysis variables, compared with the estimated value generated the most suitable anisotropic semivariogram models for the data is anisotropic semivariogram gauss value with Mean Square Error of 0.26 × 107.