/////////////////////////////////////////// // Running convGAN-majority-5 on imblearn_mammography /////////////////////////////////////////// Load 'data_input/imblearn_mammography' from imblearn non empty cut in data_input/imblearn_mammography! (7 points) Data loaded. -> Shuffling data ### Start exercise for synthetic point generator ====== Step 1/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 1/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2126, 59 GAN fn, tp: 7, 45 GAN f1 score: 0.577 GAN cohens kappa score: 0.563 -> test with 'LR' LR tn, fp: 1831, 354 LR fn, tp: 6, 46 LR f1 score: 0.204 LR cohens kappa score: 0.169 LR average precision score: 0.560 -> test with 'GB' GB tn, fp: 2123, 62 GB fn, tp: 16, 36 GB f1 score: 0.480 GB cohens kappa score: 0.464 -> test with 'KNN' KNN tn, fp: 2092, 93 KNN fn, tp: 6, 46 KNN f1 score: 0.482 KNN cohens kappa score: 0.464 ------ Step 1/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2142, 43 GAN fn, tp: 13, 39 GAN f1 score: 0.582 GAN cohens kappa score: 0.570 -> test with 'LR' LR tn, fp: 1909, 276 LR fn, tp: 6, 46 LR f1 score: 0.246 LR cohens kappa score: 0.215 LR average precision score: 0.489 -> test with 'GB' GB tn, fp: 2140, 45 GB fn, tp: 12, 40 GB f1 score: 0.584 GB cohens kappa score: 0.572 -> test with 'KNN' KNN tn, fp: 2109, 76 KNN fn, tp: 8, 44 KNN f1 score: 0.512 KNN cohens kappa score: 0.495 ------ Step 1/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2128, 57 GAN fn, tp: 7, 45 GAN f1 score: 0.584 GAN cohens kappa score: 0.571 -> test with 'LR' LR tn, fp: 1907, 278 LR fn, tp: 6, 46 LR f1 score: 0.245 LR cohens kappa score: 0.213 LR average precision score: 0.590 -> test with 'GB' GB tn, fp: 2131, 54 GB fn, tp: 9, 43 GB f1 score: 0.577 GB cohens kappa score: 0.564 -> test with 'KNN' KNN tn, fp: 2105, 80 KNN fn, tp: 7, 45 KNN f1 score: 0.508 KNN cohens kappa score: 0.492 ------ Step 1/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2085, 100 GAN fn, tp: 10, 42 GAN f1 score: 0.433 GAN cohens kappa score: 0.413 -> test with 'LR' LR tn, fp: 1887, 298 LR fn, tp: 6, 46 LR f1 score: 0.232 LR cohens kappa score: 0.200 LR average precision score: 0.327 -> test with 'GB' GB tn, fp: 2142, 43 GB fn, tp: 17, 35 GB f1 score: 0.538 GB cohens kappa score: 0.525 -> test with 'KNN' KNN tn, fp: 2089, 96 KNN fn, tp: 9, 43 KNN f1 score: 0.450 KNN cohens kappa score: 0.431 ------ Step 1/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8532 synthetic samples -> test with GAN.predict GAN tn, fp: 2111, 72 GAN fn, tp: 9, 43 GAN f1 score: 0.515 GAN cohens kappa score: 0.499 -> test with 'LR' LR tn, fp: 1914, 269 LR fn, tp: 5, 47 LR f1 score: 0.255 LR cohens kappa score: 0.224 LR average precision score: 0.568 -> test with 'GB' GB tn, fp: 2139, 44 GB fn, tp: 13, 39 GB f1 score: 0.578 GB cohens kappa score: 0.565 -> test with 'KNN' KNN tn, fp: 2093, 90 KNN fn, tp: 11, 41 KNN f1 score: 0.448 KNN cohens kappa score: 0.429 ====== Step 2/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 2/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2097, 88 GAN fn, tp: 10, 42 GAN f1 score: 0.462 GAN cohens kappa score: 0.443 -> test with 'LR' LR tn, fp: 1879, 306 LR fn, tp: 6, 46 LR f1 score: 0.228 LR cohens kappa score: 0.195 LR average precision score: 0.487 -> test with 'GB' GB tn, fp: 2115, 70 GB fn, tp: 12, 40 GB f1 score: 0.494 GB cohens kappa score: 0.477 -> test with 'KNN' KNN tn, fp: 2078, 107 KNN fn, tp: 8, 44 KNN f1 score: 0.433 KNN cohens kappa score: 0.413 ------ Step 2/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2077, 108 GAN fn, tp: 3, 49 GAN f1 score: 0.469 GAN cohens kappa score: 0.450 -> test with 'LR' LR tn, fp: 1878, 307 LR fn, tp: 8, 44 LR f1 score: 0.218 LR cohens kappa score: 0.185 LR average precision score: 0.454 -> test with 'GB' GB tn, fp: 2125, 60 GB fn, tp: 13, 39 GB f1 score: 0.517 GB cohens kappa score: 0.501 -> test with 'KNN' KNN tn, fp: 2079, 106 KNN fn, tp: 7, 45 KNN f1 score: 0.443 KNN cohens kappa score: 0.423 ------ Step 2/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2149, 36 GAN fn, tp: 12, 40 GAN f1 score: 0.625 GAN cohens kappa score: 0.614 -> test with 'LR' LR tn, fp: 1805, 380 LR fn, tp: 7, 45 LR f1 score: 0.189 LR cohens kappa score: 0.154 LR average precision score: 0.512 -> test with 'GB' GB tn, fp: 2143, 42 GB fn, tp: 12, 40 GB f1 score: 0.597 GB cohens kappa score: 0.585 -> test with 'KNN' KNN tn, fp: 2104, 81 KNN fn, tp: 10, 42 KNN f1 score: 0.480 KNN cohens kappa score: 0.462 ------ Step 2/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2094, 91 GAN fn, tp: 5, 47 GAN f1 score: 0.495 GAN cohens kappa score: 0.477 -> test with 'LR' LR tn, fp: 1816, 369 LR fn, tp: 3, 49 LR f1 score: 0.209 LR cohens kappa score: 0.174 LR average precision score: 0.505 -> test with 'GB' GB tn, fp: 2133, 52 GB fn, tp: 11, 41 GB f1 score: 0.566 GB cohens kappa score: 0.552 -> test with 'KNN' KNN tn, fp: 2088, 97 KNN fn, tp: 6, 46 KNN f1 score: 0.472 KNN cohens kappa score: 0.453 ------ Step 2/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8532 synthetic samples -> test with GAN.predict GAN tn, fp: 2135, 48 GAN fn, tp: 11, 41 GAN f1 score: 0.582 GAN cohens kappa score: 0.569 -> test with 'LR' LR tn, fp: 1925, 258 LR fn, tp: 8, 44 LR f1 score: 0.249 LR cohens kappa score: 0.218 LR average precision score: 0.504 -> test with 'GB' GB tn, fp: 2156, 27 GB fn, tp: 16, 36 GB f1 score: 0.626 GB cohens kappa score: 0.616 -> test with 'KNN' KNN tn, fp: 2113, 70 KNN fn, tp: 11, 41 KNN f1 score: 0.503 KNN cohens kappa score: 0.487 ====== Step 3/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 3/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2111, 74 GAN fn, tp: 5, 47 GAN f1 score: 0.543 GAN cohens kappa score: 0.528 -> test with 'LR' LR tn, fp: 1888, 297 LR fn, tp: 6, 46 LR f1 score: 0.233 LR cohens kappa score: 0.201 LR average precision score: 0.543 -> test with 'GB' GB tn, fp: 2149, 36 GB fn, tp: 10, 42 GB f1 score: 0.646 GB cohens kappa score: 0.636 -> test with 'KNN' KNN tn, fp: 2102, 83 KNN fn, tp: 6, 46 KNN f1 score: 0.508 KNN cohens kappa score: 0.491 ------ Step 3/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2063, 122 GAN fn, tp: 8, 44 GAN f1 score: 0.404 GAN cohens kappa score: 0.382 -> test with 'LR' LR tn, fp: 1906, 279 LR fn, tp: 7, 45 LR f1 score: 0.239 LR cohens kappa score: 0.208 LR average precision score: 0.412 -> test with 'GB' GB tn, fp: 2140, 45 GB fn, tp: 17, 35 GB f1 score: 0.530 GB cohens kappa score: 0.517 -> test with 'KNN' KNN tn, fp: 2092, 93 KNN fn, tp: 10, 42 KNN f1 score: 0.449 KNN cohens kappa score: 0.430 ------ Step 3/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2104, 81 GAN fn, tp: 3, 49 GAN f1 score: 0.538 GAN cohens kappa score: 0.523 -> test with 'LR' LR tn, fp: 1889, 296 LR fn, tp: 3, 49 LR f1 score: 0.247 LR cohens kappa score: 0.215 LR average precision score: 0.488 -> test with 'GB' GB tn, fp: 2134, 51 GB fn, tp: 10, 42 GB f1 score: 0.579 GB cohens kappa score: 0.566 -> test with 'KNN' KNN tn, fp: 2101, 84 KNN fn, tp: 4, 48 KNN f1 score: 0.522 KNN cohens kappa score: 0.505 ------ Step 3/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2124, 61 GAN fn, tp: 12, 40 GAN f1 score: 0.523 GAN cohens kappa score: 0.508 -> test with 'LR' LR tn, fp: 1892, 293 LR fn, tp: 9, 43 LR f1 score: 0.222 LR cohens kappa score: 0.189 LR average precision score: 0.479 -> test with 'GB' GB tn, fp: 2143, 42 GB fn, tp: 16, 36 GB f1 score: 0.554 GB cohens kappa score: 0.541 -> test with 'KNN' KNN tn, fp: 2101, 84 KNN fn, tp: 14, 38 KNN f1 score: 0.437 KNN cohens kappa score: 0.418 ------ Step 3/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8532 synthetic samples -> test with GAN.predict GAN tn, fp: 2111, 72 GAN fn, tp: 14, 38 GAN f1 score: 0.469 GAN cohens kappa score: 0.452 -> test with 'LR' LR tn, fp: 1906, 277 LR fn, tp: 7, 45 LR f1 score: 0.241 LR cohens kappa score: 0.209 LR average precision score: 0.567 -> test with 'GB' GB tn, fp: 2144, 39 GB fn, tp: 15, 37 GB f1 score: 0.578 GB cohens kappa score: 0.566 -> test with 'KNN' KNN tn, fp: 2106, 77 KNN fn, tp: 11, 41 KNN f1 score: 0.482 KNN cohens kappa score: 0.465 ====== Step 4/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 4/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2104, 81 GAN fn, tp: 10, 42 GAN f1 score: 0.480 GAN cohens kappa score: 0.462 -> test with 'LR' LR tn, fp: 1918, 267 LR fn, tp: 7, 45 LR f1 score: 0.247 LR cohens kappa score: 0.216 LR average precision score: 0.564 -> test with 'GB' GB tn, fp: 2142, 43 GB fn, tp: 18, 34 GB f1 score: 0.527 GB cohens kappa score: 0.514 -> test with 'KNN' KNN tn, fp: 2120, 65 KNN fn, tp: 10, 42 KNN f1 score: 0.528 KNN cohens kappa score: 0.513 ------ Step 4/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2104, 81 GAN fn, tp: 6, 46 GAN f1 score: 0.514 GAN cohens kappa score: 0.497 -> test with 'LR' LR tn, fp: 1898, 287 LR fn, tp: 5, 47 LR f1 score: 0.244 LR cohens kappa score: 0.212 LR average precision score: 0.401 -> test with 'GB' GB tn, fp: 2139, 46 GB fn, tp: 17, 35 GB f1 score: 0.526 GB cohens kappa score: 0.513 -> test with 'KNN' KNN tn, fp: 1466, 719 KNN fn, tp: 5, 47 KNN f1 score: 0.115 KNN cohens kappa score: 0.075 ------ Step 4/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2120, 65 GAN fn, tp: 9, 43 GAN f1 score: 0.537 GAN cohens kappa score: 0.523 -> test with 'LR' LR tn, fp: 1901, 284 LR fn, tp: 7, 45 LR f1 score: 0.236 LR cohens kappa score: 0.204 LR average precision score: 0.468 -> test with 'GB' GB tn, fp: 2135, 50 GB fn, tp: 8, 44 GB f1 score: 0.603 GB cohens kappa score: 0.590 -> test with 'KNN' KNN tn, fp: 2095, 90 KNN fn, tp: 8, 44 KNN f1 score: 0.473 KNN cohens kappa score: 0.455 ------ Step 4/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2081, 104 GAN fn, tp: 6, 46 GAN f1 score: 0.455 GAN cohens kappa score: 0.436 -> test with 'LR' LR tn, fp: 1917, 268 LR fn, tp: 9, 43 LR f1 score: 0.237 LR cohens kappa score: 0.205 LR average precision score: 0.484 -> test with 'GB' GB tn, fp: 2133, 52 GB fn, tp: 12, 40 GB f1 score: 0.556 GB cohens kappa score: 0.542 -> test with 'KNN' KNN tn, fp: 2084, 101 KNN fn, tp: 7, 45 KNN f1 score: 0.455 KNN cohens kappa score: 0.435 ------ Step 4/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8532 synthetic samples -> test with GAN.predict GAN tn, fp: 2125, 58 GAN fn, tp: 9, 43 GAN f1 score: 0.562 GAN cohens kappa score: 0.548 -> test with 'LR' LR tn, fp: 1889, 294 LR fn, tp: 1, 51 LR f1 score: 0.257 LR cohens kappa score: 0.226 LR average precision score: 0.465 -> test with 'GB' GB tn, fp: 2143, 40 GB fn, tp: 12, 40 GB f1 score: 0.606 GB cohens kappa score: 0.595 -> test with 'KNN' KNN tn, fp: 2087, 96 KNN fn, tp: 9, 43 KNN f1 score: 0.450 KNN cohens kappa score: 0.431 ====== Step 5/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 5/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2121, 64 GAN fn, tp: 6, 46 GAN f1 score: 0.568 GAN cohens kappa score: 0.554 -> test with 'LR' LR tn, fp: 1890, 295 LR fn, tp: 4, 48 LR f1 score: 0.243 LR cohens kappa score: 0.211 LR average precision score: 0.493 -> test with 'GB' GB tn, fp: 2145, 40 GB fn, tp: 12, 40 GB f1 score: 0.606 GB cohens kappa score: 0.595 -> test with 'KNN' KNN tn, fp: 1444, 741 KNN fn, tp: 7, 45 KNN f1 score: 0.107 KNN cohens kappa score: 0.067 ------ Step 5/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2111, 74 GAN fn, tp: 8, 44 GAN f1 score: 0.518 GAN cohens kappa score: 0.502 -> test with 'LR' LR tn, fp: 1898, 287 LR fn, tp: 6, 46 LR f1 score: 0.239 LR cohens kappa score: 0.207 LR average precision score: 0.422 -> test with 'GB' GB tn, fp: 2137, 48 GB fn, tp: 14, 38 GB f1 score: 0.551 GB cohens kappa score: 0.537 -> test with 'KNN' KNN tn, fp: 1417, 768 KNN fn, tp: 6, 46 KNN f1 score: 0.106 KNN cohens kappa score: 0.065 ------ Step 5/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2130, 55 GAN fn, tp: 13, 39 GAN f1 score: 0.534 GAN cohens kappa score: 0.520 -> test with 'LR' LR tn, fp: 1857, 328 LR fn, tp: 8, 44 LR f1 score: 0.208 LR cohens kappa score: 0.174 LR average precision score: 0.531 -> test with 'GB' GB tn, fp: 2133, 52 GB fn, tp: 19, 33 GB f1 score: 0.482 GB cohens kappa score: 0.466 -> test with 'KNN' KNN tn, fp: 2096, 89 KNN fn, tp: 11, 41 KNN f1 score: 0.451 KNN cohens kappa score: 0.432 ------ Step 5/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8530 synthetic samples -> test with GAN.predict GAN tn, fp: 2118, 67 GAN fn, tp: 8, 44 GAN f1 score: 0.540 GAN cohens kappa score: 0.525 -> test with 'LR' LR tn, fp: 1891, 294 LR fn, tp: 4, 48 LR f1 score: 0.244 LR cohens kappa score: 0.212 LR average precision score: 0.477 -> test with 'GB' GB tn, fp: 2130, 55 GB fn, tp: 11, 41 GB f1 score: 0.554 GB cohens kappa score: 0.540 -> test with 'KNN' KNN tn, fp: 2095, 90 KNN fn, tp: 8, 44 KNN f1 score: 0.473 KNN cohens kappa score: 0.455 ------ Step 5/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 8532 synthetic samples -> test with GAN.predict GAN tn, fp: 2056, 127 GAN fn, tp: 9, 43 GAN f1 score: 0.387 GAN cohens kappa score: 0.365 -> test with 'LR' LR tn, fp: 1930, 253 LR fn, tp: 8, 44 LR f1 score: 0.252 LR cohens kappa score: 0.221 LR average precision score: 0.576 -> test with 'GB' GB tn, fp: 2134, 49 GB fn, tp: 14, 38 GB f1 score: 0.547 GB cohens kappa score: 0.533 -> test with 'KNN' KNN tn, fp: 2094, 89 KNN fn, tp: 11, 41 KNN f1 score: 0.451 KNN cohens kappa score: 0.432 ### Exercise is done. -----[ LR ]----- maximum: LR tn, fp: 1930, 380 LR fn, tp: 9, 51 LR f1 score: 0.257 LR cohens kappa score: 0.226 LR average precision score: 0.590 average: LR tn, fp: 1888.84, 295.76 LR fn, tp: 6.08, 45.92 LR f1 score: 0.234 LR cohens kappa score: 0.202 LR average precision score: 0.495 minimum: LR tn, fp: 1805, 253 LR fn, tp: 1, 43 LR f1 score: 0.189 LR cohens kappa score: 0.154 LR average precision score: 0.327 -----[ GB ]----- maximum: GB tn, fp: 2156, 70 GB fn, tp: 19, 44 GB f1 score: 0.646 GB cohens kappa score: 0.636 average: GB tn, fp: 2137.12, 47.48 GB fn, tp: 13.44, 38.56 GB f1 score: 0.560 GB cohens kappa score: 0.547 minimum: GB tn, fp: 2115, 27 GB fn, tp: 8, 33 GB f1 score: 0.480 GB cohens kappa score: 0.464 -----[ KNN ]----- maximum: KNN tn, fp: 2120, 768 KNN fn, tp: 14, 48 KNN f1 score: 0.528 KNN cohens kappa score: 0.513 average: KNN tn, fp: 2018.0, 166.6 KNN fn, tp: 8.4, 43.6 KNN f1 score: 0.430 KNN cohens kappa score: 0.409 minimum: KNN tn, fp: 1417, 65 KNN fn, tp: 4, 38 KNN f1 score: 0.106 KNN cohens kappa score: 0.065 -----[ GAN ]----- maximum: GAN tn, fp: 2149, 127 GAN fn, tp: 14, 49 GAN f1 score: 0.625 GAN cohens kappa score: 0.614 average: GAN tn, fp: 2109.08, 75.52 GAN fn, tp: 8.52, 43.48 GAN f1 score: 0.516 GAN cohens kappa score: 0.500 minimum: GAN tn, fp: 2056, 36 GAN fn, tp: 3, 38 GAN f1 score: 0.387 GAN cohens kappa score: 0.365