API tools faq
paste
Bananaware

quality evaluation

Bananaware
Sep 12th, 2022
57
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 6.36 KB | None | 0 0
raw download clone embed print report
  1. datasets:
  2.  
  3. luiz: 147 Pos | 180 Neg (subset jersey)
  4. jersey: 749 Pos | 702 Neg
  5. usp: 694 Pos | 292 Neg
  6. puruna: 210 Pos | 3543 Neg
  7.  
  8.  
  9. luiz: doc de avaliação de qualidade dos grafos. qualidade [good, bad, terrible]
  10. outras: avaliação visual de qualidade feita por mim e nossos grandes antepassados. qualidade [good, medium, bad, artifact]
  11.  
  12.  
  13.  
  14. --------- big trees (depth = 12~16) ---------
  15.  
  16. train: luiz | https://i.imgur.com/3qOd2js.png
  17.  
  18. test: jersey | Dataset: 749 Pos | 702 Neg
  19. Predict: 262 Pos | 1189 Neg
  20. True Pos: 77 | True Neg: 517 | False Pos: 185 | False Neg: 672
  21. Accuracy: 0.41 | Precision: 0.29 | Recall: 0.10 | Specificity: 0.74 | F1: 0.15
  22.  
  23. test: usp | Dataset: 694 Pos | 292 Neg
  24. Predict: 437 Pos | 549 Neg
  25. True Pos: 278 | True Neg: 133 | False Pos: 159 | False Neg: 416
  26. Accuracy: 0.42 | Precision: 0.64 | Recall: 0.40 | Specificity: 0.46 | F1: 0.49
  27.  
  28. test: puruna | Dataset: 210 Pos | 3543 Neg
  29. Predict: 1548 Pos | 2205 Neg
  30. True Pos: 27 | True Neg: 2022 | False Pos: 1521 | False Neg: 183
  31. Accuracy: 0.55 | Precision: 0.02 | Recall: 0.13 | Specificity: 0.57 | F1: 0.03
  32.  
  33.  
  34. ==
  35.  
  36.  
  37. train: jersey | https://i.imgur.com/UyS4IA6.png
  38.  
  39. test: usp | Dataset: 694 Pos | 292 Neg
  40. Predict: 109 Pos | 877 Neg
  41. True Pos: 98 | True Neg: 281 | False Pos: 11 | False Neg: 596
  42. Accuracy: 0.38 | Precision: 0.90 | Recall: 0.14 | Specificity: 0.96 | F1: 0.24
  43.  
  44. test: puruna | Dataset: 210 Pos | 3543 Neg
  45. Predict: 220 Pos | 3533 Neg
  46. True Pos: 107 | True Neg: 3430 | False Pos: 113 | False Neg: 103
  47. Accuracy: 0.94 | Precision: 0.49 | Recall: 0.51 | Specificity: 0.97 | F1: 0.50
  48.  
  49. test: luiz | Dataset: 147 Pos | 180 Neg
  50. Predict: 327 Pos | 0 Neg
  51. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  52. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  53.  
  54.  
  55. ==
  56.  
  57.  
  58. train: usp | https://i.imgur.com/L2DEXdd.png
  59.  
  60. test: jersey | Dataset: 749 Pos | 702 Neg
  61. Predict: 981 Pos | 470 Neg
  62. True Pos: 633 | True Neg: 354 | False Pos: 348 | False Neg: 116
  63. Accuracy: 0.68 | Precision: 0.65 | Recall: 0.85 | Specificity: 0.50 | F1: 0.73
  64.  
  65. test: puruna | Dataset: 210 Pos | 3543 Neg
  66. Predict: 803 Pos | 2950 Neg
  67. True Pos: 144 | True Neg: 2884 | False Pos: 659 | False Neg: 66
  68. Accuracy: 0.81 | Precision: 0.18 | Recall: 0.69 | Specificity: 0.81 | F1: 0.28
  69.  
  70. test: luiz | Dataset: 147 Pos | 180 Neg
  71. Predict: 327 Pos | 0 Neg
  72. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  73. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  74.  
  75.  
  76. ==
  77.  
  78.  
  79. train: puruna | https://i.imgur.com/nFYKmej.png
  80.  
  81. test: jersey | Dataset: 749 Pos | 702 Neg
  82. Predict: 278 Pos | 1173 Neg
  83. True Pos: 253 | True Neg: 677 | False Pos: 25 | False Neg: 496
  84. Accuracy: 0.64 | Precision: 0.91 | Recall: 0.34 | Specificity: 0.96 | F1: 0.49
  85.  
  86. test: usp | Dataset: 694 Pos | 292 Neg
  87. Predict: 15 Pos | 971 Neg
  88. True Pos: 14 | True Neg: 291 | False Pos: 1 | False Neg: 680
  89. Accuracy: 0.31 | Precision: 0.93 | Recall: 0.02 | Specificity: 1.00 | F1: 0.04
  90.  
  91. test: luiz | Dataset: 147 Pos | 180 Neg
  92. Predict: 327 Pos | 0 Neg
  93. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  94. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  95.  
  96.  
  97.  
  98.  
  99. --------- small trees (depth = 3) ---------
  100.  
  101. train: luiz | https://i.imgur.com/UfhrdeP.png
  102.  
  103. test: jersey | Dataset: 749 Pos | 702 Neg
  104. Predict: 206 Pos | 1245 Neg
  105. True Pos: 56 | True Neg: 552 | False Pos: 150 | False Neg: 693
  106. Accuracy: 0.42 | Precision: 0.27 | Recall: 0.07 | Specificity: 0.79 | F1: 0.12
  107.  
  108. test: usp | Dataset: 694 Pos | 292 Neg
  109. Predict: 362 Pos | 624 Neg
  110. True Pos: 235 | True Neg: 165 | False Pos: 127 | False Neg: 459
  111. Accuracy: 0.41 | Precision: 0.65 | Recall: 0.34 | Specificity: 0.57 | F1: 0.45
  112.  
  113. test: puruna | Dataset: 210 Pos | 3543 Neg
  114. Predict: 1460 Pos | 2293 Neg
  115. True Pos: 18 | True Neg: 2101 | False Pos: 1442 | False Neg: 192
  116. Accuracy: 0.56 | Precision: 0.01 | Recall: 0.09 | Specificity: 0.59 | F1: 0.02
  117.  
  118.  
  119. ==
  120.  
  121.  
  122. train: jersey | https://i.imgur.com/cn8ZI24.png
  123.  
  124. test: usp | Dataset: 694 Pos | 292 Neg
  125. Predict: 164 Pos | 822 Neg
  126. True Pos: 112 | True Neg: 280 | False Pos: 12 | False Neg: 582
  127. Accuracy: 0.40 | Precision: 0.90 | Recall: 0.16 | Specificity: 0.96 | F1: 0.27
  128. [avg: -0.391, med: -0.534, min: -0.892, max: 0.739]
  129.  
  130. test: puruna | Dataset: 210 Pos | 3543 Neg
  131. Predict: 220 Pos | 3533 Neg
  132. True Pos: 115 | True Neg: 3373 | False Pos: 170 | False Neg: 95
  133. Accuracy: 0.93 | Precision: 0.40 | Recall: 0.55 | Specificity: 0.95 | F1: 0.46
  134. [avg: -0.630, med: -0.753, min: -0.901, max: 0.747]
  135.  
  136. test: luiz | Dataset: 147 Pos | 180 Neg
  137. Predict: 327 Pos | 0 Neg
  138. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  139. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  140. [0.74297833 0.74528952 0.74421003 0.74100156 0.74401376 ... 0.74061509]
  141. [avg: 0.741, med: 0.742, min: 0.657, max: 0.746]
  142.  
  143.  
  144. ==
  145.  
  146.  
  147. train: usp | https://i.imgur.com/kj8bqv1.png
  148.  
  149. test: jersey | Dataset: 749 Pos | 702 Neg
  150. Predict: 1324 Pos | 127 Neg
  151. True Pos: 735 | True Neg: 113 | False Pos: 589 | False Neg: 14
  152. Accuracy: 0.58 | Precision: 0.56 | Recall: 0.98 | Specificity: 0.16 | F1: 0.71
  153.  
  154. test: puruna | Dataset: 210 Pos | 3543 Neg
  155. Predict: 2721 Pos | 1032 Neg
  156. True Pos: 194 | True Neg: 1016 | False Pos: 2527 | False Neg: 16
  157. Accuracy: 0.32 | Precision: 0.07 | Recall: 0.92 | Specificity: 0.29 | F1: 0.13
  158.  
  159. test: luiz | Dataset: 147 Pos | 180 Neg
  160. Predict: 327 Pos | 0 Neg
  161. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  162. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  163.  
  164.  
  165. ==
  166.  
  167.  
  168. train: puruna | https://i.imgur.com/t7oWCnq.png
  169.  
  170. test: jersey | Dataset: 749 Pos | 702 Neg
  171. Predict: 372 Pos | 1079 Neg
  172. True Pos: 327 | True Neg: 657 | False Pos: 45 | False Neg: 422
  173. Accuracy: 0.68 | Precision: 0.88 | Recall: 0.44 | Specificity: 0.94 | F1: 0.58
  174.  
  175. test: usp | Dataset: 694 Pos | 292 Neg
  176. Predict: 31 Pos | 955 Neg
  177. True Pos: 28 | True Neg: 289 | False Pos: 3 | False Neg: 666
  178. Accuracy: 0.32 | Precision: 0.90 | Recall: 0.04 | Specificity: 0.99 | F1: 0.08
  179.  
  180. test: luiz | Dataset: 147 Pos | 180 Neg
  181. Predict: 327 Pos | 0 Neg
  182. True Pos: 147 | True Neg: 0 | False Pos: 180 | False Neg: 0
  183. Accuracy: 0.45 | Precision: 0.45 | Recall: 1.00 | Specificity: 0.00 | F1: 0.62
  184.  
  185.  
  186.  
  187.  
  188. provavelmente mais relevantes:
  189. - p_neigh (especialmente p3neigh), n_edges
  190. - talvez dist_max, dist_med
  191.  
  192. thresholding...
Advertisement
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!