Data dimension: 30000 x 25 Missing values: 0 Y table: Y 0 1 23364 6636 Y proportion: Y 0 1 0.7788 0.2212 ==== Descriptive statistics ==== Variable Mean SD Min Median Max X1 X1 167484.32 129747.66 10000 140000 1000000 X2 X2 1.60 0.49 1 2 2 X3 X3 1.85 0.79 0 2 6 X4 X4 1.55 0.52 0 2 3 X5 X5 35.49 9.22 21 34 79 X6 X6 -0.02 1.12 -2 0 8 X7 X7 -0.13 1.20 -2 0 8 X8 X8 -0.17 1.20 -2 0 8 X9 X9 -0.22 1.17 -2 0 8 X10 X10 -0.27 1.13 -2 0 8 X11 X11 -0.29 1.15 -2 0 8 X12 X12 51223.33 73635.86 -165580 22382 964511 X13 X13 49179.08 71173.77 -69777 21200 983931 X14 X14 47013.15 69349.39 -157264 20088 1664089 X15 X15 43262.95 64332.86 -170000 19052 891586 X16 X16 40311.40 60797.16 -81334 18104 927171 X17 X17 38871.76 59554.11 -339603 17071 961664 X18 X18 5663.58 16563.28 0 2100 873552 X19 X19 5921.16 23040.87 0 2009 1684259 X20 X20 5225.68 17606.96 0 1800 896040 X21 X21 4826.08 15666.16 0 1500 621000 X22 X22 4799.39 15278.31 0 1500 426529 X23 X23 5215.50 17777.47 0 1500 528666 null device 1 null device 1 ==== Logistic confusion matrix ==== Predicted Actual 0 1 0 4539 134 1 980 348 ==== Logistic metrics ==== Accuracy Sensitivity Specificity Precision ErrorRate 1 0.8144 0.262 0.9713 0.722 0.1856 Logistic AUC: 0.7253 ==== Top logistic coefficients by |z| ==== Estimate Std. Error z value Pr(>|z|) X6 0.5809 0.0198 29.398 0e+00 X18 0.0000 0.0000 -6.904 0e+00 X12 0.0000 0.0000 -5.054 0e+00 X3 -0.1057 0.0234 -4.513 0e+00 X5 0.0085 0.0020 4.235 0e+00 X4 -0.1484 0.0356 -4.174 0e+00 X1 0.0000 0.0000 -4.167 0e+00 X8 0.0915 0.0254 3.608 3e-04 ==== PCA variance table ==== PC Eigenvalue Proportion Cumulative 1 PC1 6.5431 0.2845 0.2845 2 PC2 4.0983 0.1782 0.4627 3 PC3 1.5510 0.0674 0.5301 4 PC4 1.4723 0.0640 0.5941 5 PC5 1.0252 0.0446 0.6387 6 PC6 0.9572 0.0416 0.6803 7 PC7 0.9076 0.0395 0.7198 8 PC8 0.8876 0.0386 0.7584 9 PC9 0.8712 0.0379 0.7962 10 PC10 0.7829 0.0340 0.8303 PCA k eigenvalue>1: 5 PCA k cumulative>=80%: 10 ==== PCA top variables ==== PC1_positive PC1_negative PC2_positive PC2_negative 1 X15 X2 X9 X1 2 X16 X4 X8 X20 3 X14 X5 X7 X18 4 X13 X3 X10 X14 5 X17 X1 X11 X15 null device 1 null device 1 null device 1 ==== Factor analysis loadings ==== Variable Factor1 Factor2 Factor3 Factor4 Factor5 Communality Uniqueness X1 X1 0.293 -0.358 0.286 0.149 0.053 0.320 0.680 X2 X2 -0.015 -0.069 -0.013 0.020 0.014 0.006 0.994 X3 X3 0.011 0.127 -0.072 -0.079 -0.024 0.028 0.972 X4 X4 -0.032 0.048 -0.004 -0.007 -0.008 0.003 0.997 X5 X5 0.060 -0.071 0.037 0.006 0.003 0.010 0.990 X6 X6 0.140 0.619 -0.163 -0.063 -0.043 0.435 0.565 X7 X7 0.159 0.755 -0.137 -0.065 -0.039 0.621 0.379 X8 X8 0.131 0.830 -0.061 -0.039 -0.021 0.711 0.289 X9 X9 0.111 0.891 -0.016 0.004 0.069 0.812 0.188 X10 X10 0.111 0.880 0.013 0.078 0.064 0.797 0.203 X11 X11 0.130 0.806 -0.001 0.127 0.052 0.685 0.315 X12 X12 0.915 0.127 0.199 -0.086 -0.119 0.914 0.086 X13 X13 0.933 0.152 0.279 -0.081 -0.127 0.995 0.005 X14 X14 0.934 0.145 0.210 -0.071 0.231 0.995 0.005 X15 X15 0.912 0.157 0.166 0.172 0.115 0.927 0.073 X16 X16 0.909 0.153 0.095 0.359 0.079 0.995 0.005 X17 X17 0.871 0.154 0.110 0.317 0.076 0.901 0.099 X18 X18 0.126 -0.013 0.621 0.116 0.028 0.416 0.584 X19 X19 0.097 -0.053 0.384 0.031 0.668 0.607 0.393 X20 X20 0.095 -0.050 0.371 0.390 0.009 0.302 0.698 X21 X21 0.146 -0.075 0.204 0.407 0.013 0.234 0.766 X22 X22 0.121 -0.060 0.262 0.044 0.103 0.099 0.900 X23 X23 0.128 -0.052 0.278 0.062 0.068 0.105 0.895 Factor variance proportions: Factor1 Factor2 Factor3 Factor4 Factor5 0.2295 0.1811 0.0525 0.0294 0.0257 Factor cumulative variance: 0.2295 0.4106 0.4631 0.4925 0.5182 ==== Factor top variables ==== Factor1 Factor2 Factor3 1 X14, X13, X12, X15, X16 X9, X10, X8, X11, X7 X18, X19, X20, X1, X13 Factor4 Factor5 1 X21, X20, X16, X17, X15 X19, X14, X13, X12, X15 null device 1 null device 1 ==== Hierarchical clustering metrics ==== Method Accuracy Sensitivity Specificity Precision ErrorRate 1 single 0.7786 0.0000 0.9996 0.0000 0.2214 2 average 0.7786 0.0000 0.9996 0.0000 0.2214 3 complete 0.7786 0.0000 0.9996 0.0000 0.2214 4 ward.D2 0.6969 0.1554 0.8506 0.2279 0.3031 null device 1 ==== Kmeans clustering metrics ==== Algorithm Accuracy Sensitivity Specificity Precision ErrorRate WithinSS 1 Hartigan-Wong 0.6906 0.1492 0.8443 0.2140 0.3094 568474 2 Lloyd 0.6903 0.1493 0.8439 0.2137 0.3097 568474 3 Forgy 0.6903 0.1493 0.8439 0.2137 0.3097 568474 4 MacQueen 0.6903 0.1493 0.8439 0.2137 0.3097 568474 null device 1 ==== Spectral clustering metrics ==== Method Accuracy Sensitivity Specificity 1 manual normalized spectral clustering 0.6389 0.4868 0.682 Precision ErrorRate 1 0.3028 0.3611 null device 1 ==== LDA train confusion ==== Predicted Actual 0 1 0 15832 522 1 3407 1238 ==== LDA train metrics ==== Accuracy Sensitivity Specificity Precision ErrorRate 1 0.8129 0.2665 0.9681 0.7034 0.1871 ==== LDA test confusion ==== Predicted Actual 0 1 0 6770 240 1 1448 543 ==== LDA test metrics ==== Accuracy Sensitivity Specificity Precision ErrorRate 1 0.8125 0.2727 0.9658 0.6935 0.1875 null device 1 ==== QDA train confusion ==== Predicted Actual 0 1 0 6543 9811 1 817 3828 ==== QDA train metrics ==== Accuracy Sensitivity Specificity Precision ErrorRate 1 0.4939 0.8241 0.4001 0.2807 0.5061 ==== QDA test confusion ==== Predicted Actual 0 1 0 2816 4194 1 308 1683 ==== QDA test metrics ==== Accuracy Sensitivity Specificity Precision ErrorRate 1 0.4998 0.8453 0.4017 0.2864 0.5002