Some Considerations and Results of Specialized Pattern Cluster Analysis with FUNDOS (PPCLUSTF) for Marine Engine Field Data

Abstract

This paper describes the Totalization and Systematization of the following analyses on the marine engine field data (1400 samples) . They are Principal Component (PCA), Normal Probability Scale (NPS), Fundermental Original Statistiscs (FUNDOS) for Specialized Pattern Cluster Analysis (PPCLUST) .
Having only an objective variable of Man-Hour per occurrence mh_i, the five vector spaces of factor scores U with all the scores of the input variables X_i by PCA are represented by the several groups of factor score U_j which is corresponding to Z_j; U₁ by the group of input variables (X₂₄, X₃₄, X₃₅) of hardware of ship and engine, U₂ by (X₁, X₂, X₃) of man-hour informations, U₃ by (X₆, X₂₁, X₃₂, X₃₃) of software of engine plants system, U₄ by (X₄, X₅, X₇, X₈) of major maintenance and U₅ by (X₁₈, X₂₂, X₂₅, X₂₈) of hardware of engine plant system.
An additional program FUNDOS is useful for beginner of this systematic program package due to the feasible fundermental statistics of original data.