Clustering and Classification methods for Biologists


MMU logo

Principal Components Analysis

LTSN Bioscience logo

Page Outline

 

Search

[ Yahoo! ] options

Factor Analysis Self Assessment

The following data were obtained from 86 plasma donors at a Liverpool hospital. As part of a much larger data set the following variables were analysed.

Descriptive Statistics

Descriptive Statistics (n = 86 cases)
    Mean Std. Deviation
HT Height (inches!) 67.37 3.97
WT Weight (pounds!) 166.22 26.67
BMI Body Mass Index 26.50 3.72
AGE Age (years) 33.83 8.43
MCV Mean Cell Volume 90.60 4.26
WBC White Blood Cell Count 6.35 1.58

 

Significant correlations are maked with a *.

Correlation Matrix
  HT WT BMI AGE MCV
Correlation

Coefficient

WT 0.534*
BMI -0.228 0.679*
AGE 0.039 0.100 0.112
MCV -0.096 -0.160 -0.162 -0.018
WBC -0.181 -0.056 0.028 -0.006 0.318*

Make a note of the significant correlations.

A preliminary PCA identified three eigen values > 1.

Eigenvalue  1.9407  1.3174  1.1144  0.9690  0.6439  0.0146
Proportion   0.323   0.220   0.186   0.162   0.107   0.002
Cumulative   0.323   0.543   0.729   0.890   0.998   1.000
  
Variable PC1 PC2 PC3 PC4 PC5 PC6 ht 0.395 0.403 0.648 -0.021 -0.135 0.493 wt 0.671 -0.178 0.227 0.138 0.067 -0.666 bmi 0.480 -0.532 -0.344 0.154 0.182 0.558 age 0.152 -0.162 -0.020 -0.974 -0.038 -0.015 wbc -0.207 -0.600 0.345 0.087 -0.686 0.018 mcv -0.313 -0.370 0.539 -0.026 0.687 0.031

Factor Analysis (Extract only those factors with eigen values > 1).

Principal Component Factor Analysis of the Correlation Matrix

  Variable  Factor1  Factor2  Factor3  Communality
  ht          0.550    0.462    0.684        0.984
  wt          0.934   -0.204    0.240        0.972
  bmi         0.668   -0.611   -0.363        0.951
  age         0.212   -0.186   -0.021        0.080
  wbc        -0.289   -0.688    0.364        0.690
  mcv        -0.436   -0.425    0.569        0.695
  
Variance 1.9407 1.3174 1.1144 4.3725 % Var 0.323 0.220 0.186 0.729 Rotated Factor Loadings and Communalities Varimax Rotation Variable Factor1 Factor2 Factor3 Communality ht -0.015 0.988 0.092 0.984 wt 0.782 0.594 0.083 0.972 bmi 0.958 -0.159 0.088 0.951 age 0.280 0.023 -0.026 0.080 wbc 0.148 -0.153 -0.803 0.690 mcv -0.169 0.026 -0.816 0.695 Variance 1.6593 1.3791 1.3341 4.3725 % Var 0.277 0.230 0.222 0.729
Factor Score Coefficients Variable Factor1 Factor2 Factor3 ht -0.130 0.749 -0.043 wt 0.417 0.355 -0.048 bmi 0.615 -0.245 0.034 age 0.175 -0.012 -0.038 wbc 0.155 -0.041 -0.613 mcv -0.064 0.135 -0.627

Scatter plot of scores on the two components.

The sex of each case is labelled. The solid lines mark the means for the two components (mean values = 0).

Scatter of fac2_1 fac1_1 by sex. 
	Both sexes are evenly spread acroos Factor 1 scores while Males tend to be above average 
	on Factor 2.

Factor Analysis 2 : Extract 4 factors, with rotation

The same data have been re-analysed, but the default extraction rule has been replaced with an option to extract 4 factors (because the eigen value for the fourth factor was 0.969).

Principal Component Factor Analysis of the Correlation Matrix
    Unrotated Factor Loadings and Communalities
            86 cases used  1 cases contain missing values

    Variable  Factor1  Factor2  Factor3  Factor4  Communality
    ht          0.550    0.462    0.684   -0.021        0.985
    wt          0.934   -0.204    0.240    0.135        0.991
    bmi         0.668   -0.611   -0.363    0.152        0.974
    age         0.212   -0.186   -0.021   -0.959        0.999
    wbc        -0.289   -0.688    0.364    0.085        0.697
    mcv        -0.436   -0.425    0.569   -0.025        0.696

    Variance   1.9407   1.3174   1.1144   0.9690       5.3416
    % Var       0.323    0.220    0.186    0.162        0.890

    Rotated Factor Loadings and Communalities
    Varimax Rotation

    Variable  Factor1  Factor2  Factor3  Factor4  Communality
    ht         -0.028    0.987    0.098   -0.016        0.985
    wt          0.788    0.605    0.050   -0.040        0.991
    bmi         0.973   -0.146    0.045   -0.060        0.974
    age         0.061    0.021    0.006   -0.997        0.999
    wbc         0.128   -0.148   -0.811    0.023        0.697
    mcv        -0.209    0.027   -0.807   -0.016        0.696

    Variance   1.6320   1.3848   1.3237   1.0011       5.3416
    % Var       0.272    0.231    0.221    0.167        0.890

    Factor Score Coefficients
    Variable  Factor1  Factor2  Factor3  Factor4
    ht         -0.144    0.748   -0.034    0.005
    wt          0.430    0.362   -0.068    0.041
    bmi         0.639   -0.236    0.003    0.019
    age        -0.048   -0.015   -0.002   -1.004
    wbc         0.143   -0.037   -0.622    0.030
    mcv        -0.099    0.137   -0.622   -0.033