舶用機関のフィールドデータへのクラスター分析について

概要

論文の詳細を見る
This paper describes the Systematization and Totalization of the following analysis on the marine field data (1399 samples) and Patternized Cluster Analysis with Principal Component Analysis abbreviated as PPCLUST, which could deal with an infinite number of input data having a several hundred number of patterns and their frequencies of factor scores. (1) Principal Component Analysis as PCA (2) Normal Probability Scale abbreviated as NPS (3) Patternized Cluster Analysis as PCLUST, which could give the factor scores for clustering in combination of factor loadings of PCA with the input scores of NPS. Having only an objective variable of Man-Hour per occurrence, all the input variables were scored by the first stage type of NPS, the following conclusion were obtained that some input variables, the two numbers of main engine sets (X21) and propellers (X33), kinds of ship (X34) and some group variables of man-power (X1, X2, X3) were very conspicuous among the thirty five ones.
社団法人日本マリンエンジニアリング学会の論文