1 / 24

Optimal Trait Scoring for Age Estimation

Optimal Trait Scoring for Age Estimation. Lyle W. Konigsberg and Susan R. Frankenberg. Options for ordinal traits. Logit , probit or exponential transitions on log or straight scale Cumulative (common standard deviation) Unrestricted cumulative (separate standard deviations)

dexter
Download Presentation

Optimal Trait Scoring for Age Estimation

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Optimal Trait Scoring for Age Estimation Lyle W. Konigsberg and Susan R. Frankenberg

  2. Options for ordinal traits • Logit, probit or exponential transitions on log or straight scale • Cumulative (common standard deviation) • Unrestricted cumulative (separate standard deviations) • Continuation ratios (forward or backward) • Stopping rules (forward or backward) • Kernel densities • Sugeno fuzzy integral

  3. Testing the normality assumption • Johnson PA. 1996. A test of the normality assumption in the ordered probit model. Metron 54:213-221. • Glewwe P. 1997. A test of the normality assumption in the ordered probit model. Econometric Reviews 16:1-19. • Weiss AA. 1997. Specification tests in ordered logit and probit models. Econometric Reviews 16:361-391.

  4. Materials Todd scores from: 422 males (Terry Collection) 332 females (Terry Collection) 163 females (Gilbert and McKern)

  5. A little history Katz and Suchey (1986) collapsed the Todd (1920) ten phase system into a “T2” system of six stages.

  6. P-values from goodness-of-fit tests

  7. Collapsing three ordered states

  8. Collapsing four ordered states

  9. Collapsing five ordered states 1+1+1+2 = 5 1+1+3 = 5 1+2+2 = 5 1+4 = 5 2+3 = 5

  10. Forming all compositions of an integer • Form all partitions of the integer (Hindenburg’s algorithm) 2+8, 3+3+4, 2+2+3+3, 2+2+2+2+2, 1+1+1+1+1+5,…, = 10 • Form all unique permutations for each partition (Knuth’s “algorithm L”) 111115, 111151, 111511, 115111, 15111, 511111

  11. The “R” script “smoosh” > smoosh(10) # of # of Total phases ways # ways 9 9 9 8 36 45 7 84 129 6 126 255 5 126 381 4 84 465 3 36 501 2 9 510 Down to how many stages? 1:

  12. > smoosh(5) [1] # of # of Total [1] phases ways # ways [1] 4 4 4 [1] 3 6 10 [1] 2 4 14 Down to how many stages? 1: 2 [,1] [,2] [,3] [,4] [,5] [1,] 1 2 3 4 4 [2,] 1 2 3 3 4 [3,] 1 2 2 3 4 [4,] 1 1 2 3 4 [5,] 1 2 3 3 3 [6,] 1 2 2 2 3 [7,] 1 1 1 2 3 [8,] 1 2 2 3 3 [9,] 1 1 2 3 3 [10,] 1 1 2 2 3 [11,] 1 2 2 2 2 [12,] 1 1 1 1 2 [13,] 1 1 2 2 2 [14,] 1 1 1 2 2

  13. Females

  14. Males

  15. Males I, II, III, IV, V, VI, VII, VIII-X

  16. Females I, II, III, IV, V, VI, VII, VIII-X

  17. Males & Females I, II, III, IV, V, VI, VII, VIII-X

  18. Females

  19. Males

  20. Males & Females

  21. Moorrees, Fanning and Hunt (1963)? > smoosh(14) [1] # of # of Total [1] phases ways # ways [1] 13 13 13 [1] 12 78 91 [1] 11 286 377 [1] 10 715 1092 [1] 9 1287 2379 [1] 8 1716 4095 [1] 7 1716 5811 [1] 6 1287 7098 [1] 5 715 7813 [1] 4 286 8099 [1] 3 78 8177 [1] 2 13 8190

  22. Some comments about “smooshing” • Not possible to “un-smoosh” data that is already “smooshed” (e.g., from Suchey-Brooks to Todd or Demirjian et al. to Moorrees, Fanning and Hunt). • The specification test provides goodness-of-fit to normal or log normal transitions. • If the fit is poor, stages can be “smooshed” until the fit is adequate. • For the Todd phases, “smooshing” showed that phases I, II, III, IV, V, VI, VII, and VIII-X fit to log normal transitions with a common log variance.

  23. Acknowledgment Data collection supported by NSF SBR-9727386 to LWK

More Related