Noi/NOTES/2012-6-13

From ZhangLabWiki
Revision as of 07:07, 14 June 2012 by >Noi (Created page with "= Regression analysis of UCLA SZ data (additional different age cutoff) = * Suggestions from lab meeting on 2012.05.16 ** LDA: random choose samples for training and (have UCL...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Regression analysis of UCLA SZ data (additional different age cutoff)

  • Suggestions from lab meeting on 2012.05.16
    • LDA: random choose samples for training and (have UCLA-SZ and UCSD-GA on training and testing data sets)
    • LDA: Change age cutoff from 40 to different number (50, 60 or 70) and see the accuracy of classification.
    • PCA: I also performed PCA and changed the color of each sample in scatter plot based on the age cutoff.
  • Training and testing data set
    • list 1-96 =>UCSD-SZ, 97-169 => UCSD-GA
    • trainAge <- c(1:40, 91:132)
    • testAge <- c(41:90, 133:169)

Tables of raw data

Age cutoff 40 (10% FDR) Predicted\Truth Adult Young adult Correct FALSE Total
Training Adult 54 2 77 5 82
Young adult 3 23
Testing Adult 59 1 85 2 87
Young adult 1 26
Cross validation Adult 112 3 161 8 169
Young adult 5 49


Age cutoff 50 (10% FDR) Predicted\Truth Adult Young adult Correct FALSE Total
Training Adult 51 4 76 6 82
Young adult 2 25
Testing Adult 58 0 85 2 87
Young adult 2 27
Cross validation Adult 110 4 162 7 169
Young adult 3 52


Age cutoff 60 (10% FDR) Predicted\Truth Adult Young adult Correct FALSE Total
Training Adult 32 7 72 10 82
Young adult 3 40
Testing Adult 40 13 71 16 87
Young adult 3 31
Cross validation Adult 72 22 141 28 169


Age cutoff 70 (10% FDR) Predicted\Truth Adult Young adult Correct FALSE Total
Training Adult 14 4 70 12 82
Young adult 8 56
Testing Adult 16 8 68 19 87
Young adult 11 52
Cross validation Adult 30 12 138 31 169
Young adult 19 108


Table of accuracy and misclassification rate

Age cutoff Training (Accuracy of allocation) Misclassicication rate Testing (quality of model) Misclassicication rate Cross validation Misclassicication rate
40 77 of 82 6.10% 85 of 87 2.30% 161 of 169 4.73%
50 76 of 82 7.32% 85 of 87 2.30% 162 of 169 4.14%
60 72 of 82 12.20% 71 of 87 18.39% 141 of 169 16.57%
70 70 of 82 14.63% 68 of 87 21.84% 138 of 169 18.34%

PCA plots File:UCLA-GA age-cutoff40 10FDR pcaplot1.png File:UCLA-GA age-cutoff40 10FDR pcaplot2.png