The particulate matter samples were studied in the surrounding of the different mills in Penang and Kedah State, Malaysia.
Multivariate statistical techniques such as multivariate analysis of variance (MANOVA), discriminant analysis (DA) and
principal components analysis (PCA) were applied in analyzing the air quality data. MANOVA showed a strong significant
difference between the five palm oil mills. DA identified two functions responsible for discriminating the mills and it showed
that the differences between mills are mainly due to total particulates, PM 1, and PM 10, affording 58% correct assignation. PCA
identified only one component responsible in explaining 93.69% of the total variance in the data representing the average of
selected parameters.