/////////////////////////////////////////// // Running convGAN-majority-5 on folding_winequality-red-4 /////////////////////////////////////////// Load 'data_input/folding_winequality-red-4' from pickle file 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 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 149, 161 GAN fn, tp: 2, 9 GAN f1 score: 0.099 GAN cohens kappa score: 0.037 -> test with 'LR' LR tn, fp: 205, 105 LR fn, tp: 5, 6 LR f1 score: 0.098 LR cohens kappa score: 0.038 LR average precision score: 0.103 -> test with 'GB' GB tn, fp: 285, 25 GB fn, tp: 10, 1 GB f1 score: 0.054 GB cohens kappa score: 0.006 -> test with 'KNN' KNN tn, fp: 209, 101 KNN fn, tp: 8, 3 KNN f1 score: 0.052 KNN cohens kappa score: -0.010 ------ Step 1/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 275, 35 GAN fn, tp: 7, 4 GAN f1 score: 0.160 GAN cohens kappa score: 0.113 -> test with 'LR' LR tn, fp: 228, 82 LR fn, tp: 4, 7 LR f1 score: 0.140 LR cohens kappa score: 0.084 LR average precision score: 0.093 -> test with 'GB' GB tn, fp: 293, 17 GB fn, tp: 6, 5 GB f1 score: 0.303 GB cohens kappa score: 0.270 -> test with 'KNN' KNN tn, fp: 241, 69 KNN fn, tp: 6, 5 KNN f1 score: 0.118 KNN cohens kappa score: 0.062 ------ Step 1/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 163, 147 GAN fn, tp: 3, 8 GAN f1 score: 0.096 GAN cohens kappa score: 0.035 -> test with 'LR' LR tn, fp: 200, 110 LR fn, tp: 3, 8 LR f1 score: 0.124 LR cohens kappa score: 0.065 LR average precision score: 0.199 -> test with 'GB' GB tn, fp: 293, 17 GB fn, tp: 8, 3 GB f1 score: 0.194 GB cohens kappa score: 0.156 -> test with 'KNN' KNN tn, fp: 220, 90 KNN fn, tp: 8, 3 KNN f1 score: 0.058 KNN cohens kappa score: -0.004 ------ Step 1/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 199, 111 GAN fn, tp: 7, 4 GAN f1 score: 0.063 GAN cohens kappa score: 0.001 -> test with 'LR' LR tn, fp: 240, 70 LR fn, tp: 5, 6 LR f1 score: 0.138 LR cohens kappa score: 0.083 LR average precision score: 0.129 -> test with 'GB' GB tn, fp: 286, 24 GB fn, tp: 10, 1 GB f1 score: 0.056 GB cohens kappa score: 0.008 -> test with 'KNN' KNN tn, fp: 226, 84 KNN fn, tp: 6, 5 KNN f1 score: 0.100 KNN cohens kappa score: 0.042 ------ Step 1/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1196 synthetic samples -> test with GAN.predict GAN tn, fp: 168, 138 GAN fn, tp: 5, 4 GAN f1 score: 0.053 GAN cohens kappa score: -0.001 -> test with 'LR' LR tn, fp: 222, 84 LR fn, tp: 4, 5 LR f1 score: 0.102 LR cohens kappa score: 0.053 LR average precision score: 0.225 -> test with 'GB' GB tn, fp: 290, 16 GB fn, tp: 9, 0 GB f1 score: 0.000 GB cohens kappa score: -0.038 -> test with 'KNN' KNN tn, fp: 234, 72 KNN fn, tp: 8, 1 KNN f1 score: 0.024 KNN cohens kappa score: -0.028 ====== Step 2/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 2/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 257, 53 GAN fn, tp: 6, 5 GAN f1 score: 0.145 GAN cohens kappa score: 0.093 -> test with 'LR' LR tn, fp: 210, 100 LR fn, tp: 1, 10 LR f1 score: 0.165 LR cohens kappa score: 0.110 LR average precision score: 0.140 -> test with 'GB' GB tn, fp: 290, 20 GB fn, tp: 8, 3 GB f1 score: 0.176 GB cohens kappa score: 0.136 -> test with 'KNN' KNN tn, fp: 216, 94 KNN fn, tp: 7, 4 KNN f1 score: 0.073 KNN cohens kappa score: 0.013 ------ Step 2/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 256, 54 GAN fn, tp: 8, 3 GAN f1 score: 0.088 GAN cohens kappa score: 0.033 -> test with 'LR' LR tn, fp: 216, 94 LR fn, tp: 3, 8 LR f1 score: 0.142 LR cohens kappa score: 0.085 LR average precision score: 0.226 -> test with 'GB' GB tn, fp: 295, 15 GB fn, tp: 11, 0 GB f1 score: 0.000 GB cohens kappa score: -0.041 -> test with 'KNN' KNN tn, fp: 238, 72 KNN fn, tp: 6, 5 KNN f1 score: 0.114 KNN cohens kappa score: 0.057 ------ Step 2/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 226, 84 GAN fn, tp: 8, 3 GAN f1 score: 0.061 GAN cohens kappa score: 0.000 -> test with 'LR' LR tn, fp: 218, 92 LR fn, tp: 3, 8 LR f1 score: 0.144 LR cohens kappa score: 0.088 LR average precision score: 0.185 -> test with 'GB' GB tn, fp: 295, 15 GB fn, tp: 9, 2 GB f1 score: 0.143 GB cohens kappa score: 0.106 -> test with 'KNN' KNN tn, fp: 234, 76 KNN fn, tp: 10, 1 KNN f1 score: 0.023 KNN cohens kappa score: -0.040 ------ Step 2/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 202, 108 GAN fn, tp: 6, 5 GAN f1 score: 0.081 GAN cohens kappa score: 0.019 -> test with 'LR' LR tn, fp: 217, 93 LR fn, tp: 5, 6 LR f1 score: 0.109 LR cohens kappa score: 0.051 LR average precision score: 0.297 -> test with 'GB' GB tn, fp: 298, 12 GB fn, tp: 8, 3 GB f1 score: 0.231 GB cohens kappa score: 0.199 -> test with 'KNN' KNN tn, fp: 240, 70 KNN fn, tp: 6, 5 KNN f1 score: 0.116 KNN cohens kappa score: 0.060 ------ Step 2/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1196 synthetic samples -> test with GAN.predict GAN tn, fp: 276, 30 GAN fn, tp: 7, 2 GAN f1 score: 0.098 GAN cohens kappa score: 0.055 -> test with 'LR' LR tn, fp: 215, 91 LR fn, tp: 3, 6 LR f1 score: 0.113 LR cohens kappa score: 0.064 LR average precision score: 0.132 -> test with 'GB' GB tn, fp: 297, 9 GB fn, tp: 7, 2 GB f1 score: 0.200 GB cohens kappa score: 0.174 -> test with 'KNN' KNN tn, fp: 232, 74 KNN fn, tp: 5, 4 KNN f1 score: 0.092 KNN cohens kappa score: 0.043 ====== Step 3/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 3/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 212, 98 GAN fn, tp: 7, 4 GAN f1 score: 0.071 GAN cohens kappa score: 0.010 -> test with 'LR' LR tn, fp: 227, 83 LR fn, tp: 4, 7 LR f1 score: 0.139 LR cohens kappa score: 0.083 LR average precision score: 0.152 -> test with 'GB' GB tn, fp: 296, 14 GB fn, tp: 10, 1 GB f1 score: 0.077 GB cohens kappa score: 0.039 -> test with 'KNN' KNN tn, fp: 227, 83 KNN fn, tp: 9, 2 KNN f1 score: 0.042 KNN cohens kappa score: -0.020 ------ Step 3/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 256, 54 GAN fn, tp: 8, 3 GAN f1 score: 0.088 GAN cohens kappa score: 0.033 -> test with 'LR' LR tn, fp: 200, 110 LR fn, tp: 2, 9 LR f1 score: 0.138 LR cohens kappa score: 0.081 LR average precision score: 0.197 -> test with 'GB' GB tn, fp: 287, 23 GB fn, tp: 9, 2 GB f1 score: 0.111 GB cohens kappa score: 0.067 -> test with 'KNN' KNN tn, fp: 234, 76 KNN fn, tp: 9, 2 KNN f1 score: 0.045 KNN cohens kappa score: -0.016 ------ Step 3/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 202, 108 GAN fn, tp: 5, 6 GAN f1 score: 0.096 GAN cohens kappa score: 0.036 -> test with 'LR' LR tn, fp: 227, 83 LR fn, tp: 5, 6 LR f1 score: 0.120 LR cohens kappa score: 0.063 LR average precision score: 0.068 -> test with 'GB' GB tn, fp: 286, 24 GB fn, tp: 9, 2 GB f1 score: 0.108 GB cohens kappa score: 0.063 -> test with 'KNN' KNN tn, fp: 226, 84 KNN fn, tp: 7, 4 KNN f1 score: 0.081 KNN cohens kappa score: 0.021 ------ Step 3/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 275, 35 GAN fn, tp: 9, 2 GAN f1 score: 0.083 GAN cohens kappa score: 0.032 -> test with 'LR' LR tn, fp: 215, 95 LR fn, tp: 3, 8 LR f1 score: 0.140 LR cohens kappa score: 0.084 LR average precision score: 0.173 -> test with 'GB' GB tn, fp: 287, 23 GB fn, tp: 9, 2 GB f1 score: 0.111 GB cohens kappa score: 0.067 -> test with 'KNN' KNN tn, fp: 228, 82 KNN fn, tp: 4, 7 KNN f1 score: 0.140 KNN cohens kappa score: 0.084 ------ Step 3/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1196 synthetic samples -> test with GAN.predict GAN tn, fp: 263, 43 GAN fn, tp: 5, 4 GAN f1 score: 0.143 GAN cohens kappa score: 0.100 -> test with 'LR' LR tn, fp: 230, 76 LR fn, tp: 3, 6 LR f1 score: 0.132 LR cohens kappa score: 0.085 LR average precision score: 0.089 -> test with 'GB' GB tn, fp: 294, 12 GB fn, tp: 6, 3 GB f1 score: 0.250 GB cohens kappa score: 0.222 -> test with 'KNN' KNN tn, fp: 239, 67 KNN fn, tp: 6, 3 KNN f1 score: 0.076 KNN cohens kappa score: 0.027 ====== Step 4/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 4/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 208, 102 GAN fn, tp: 6, 5 GAN f1 score: 0.085 GAN cohens kappa score: 0.024 -> test with 'LR' LR tn, fp: 227, 83 LR fn, tp: 3, 8 LR f1 score: 0.157 LR cohens kappa score: 0.102 LR average precision score: 0.352 -> test with 'GB' GB tn, fp: 301, 9 GB fn, tp: 8, 3 GB f1 score: 0.261 GB cohens kappa score: 0.233 -> test with 'KNN' KNN tn, fp: 240, 70 KNN fn, tp: 8, 3 KNN f1 score: 0.071 KNN cohens kappa score: 0.013 ------ Step 4/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 221, 89 GAN fn, tp: 5, 6 GAN f1 score: 0.113 GAN cohens kappa score: 0.055 -> test with 'LR' LR tn, fp: 220, 90 LR fn, tp: 3, 8 LR f1 score: 0.147 LR cohens kappa score: 0.091 LR average precision score: 0.186 -> test with 'GB' GB tn, fp: 288, 22 GB fn, tp: 9, 2 GB f1 score: 0.114 GB cohens kappa score: 0.071 -> test with 'KNN' KNN tn, fp: 209, 101 KNN fn, tp: 5, 6 KNN f1 score: 0.102 KNN cohens kappa score: 0.042 ------ Step 4/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 203, 107 GAN fn, tp: 5, 6 GAN f1 score: 0.097 GAN cohens kappa score: 0.037 -> test with 'LR' LR tn, fp: 225, 85 LR fn, tp: 5, 6 LR f1 score: 0.118 LR cohens kappa score: 0.060 LR average precision score: 0.103 -> test with 'GB' GB tn, fp: 294, 16 GB fn, tp: 9, 2 GB f1 score: 0.138 GB cohens kappa score: 0.100 -> test with 'KNN' KNN tn, fp: 219, 91 KNN fn, tp: 8, 3 KNN f1 score: 0.057 KNN cohens kappa score: -0.004 ------ Step 4/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 201, 109 GAN fn, tp: 6, 5 GAN f1 score: 0.080 GAN cohens kappa score: 0.019 -> test with 'LR' LR tn, fp: 210, 100 LR fn, tp: 1, 10 LR f1 score: 0.165 LR cohens kappa score: 0.110 LR average precision score: 0.155 -> test with 'GB' GB tn, fp: 294, 16 GB fn, tp: 8, 3 GB f1 score: 0.200 GB cohens kappa score: 0.164 -> test with 'KNN' KNN tn, fp: 230, 80 KNN fn, tp: 7, 4 KNN f1 score: 0.084 KNN cohens kappa score: 0.025 ------ Step 4/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1196 synthetic samples -> test with GAN.predict GAN tn, fp: 269, 37 GAN fn, tp: 9, 0 GAN f1 score: 0.000 GAN cohens kappa score: -0.048 -> test with 'LR' LR tn, fp: 210, 96 LR fn, tp: 6, 3 LR f1 score: 0.056 LR cohens kappa score: 0.003 LR average precision score: 0.059 -> test with 'GB' GB tn, fp: 274, 32 GB fn, tp: 8, 1 GB f1 score: 0.048 GB cohens kappa score: 0.003 -> test with 'KNN' KNN tn, fp: 230, 76 KNN fn, tp: 7, 2 KNN f1 score: 0.046 KNN cohens kappa score: -0.006 ====== Step 5/5 ======= -> Shuffling data -> Spliting data to slices ------ Step 5/5: Slice 1/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 247, 63 GAN fn, tp: 8, 3 GAN f1 score: 0.078 GAN cohens kappa score: 0.020 -> test with 'LR' LR tn, fp: 234, 76 LR fn, tp: 5, 6 LR f1 score: 0.129 LR cohens kappa score: 0.073 LR average precision score: 0.087 -> test with 'GB' GB tn, fp: 292, 18 GB fn, tp: 9, 2 GB f1 score: 0.129 GB cohens kappa score: 0.089 -> test with 'KNN' KNN tn, fp: 215, 95 KNN fn, tp: 6, 5 KNN f1 score: 0.090 KNN cohens kappa score: 0.030 ------ Step 5/5: Slice 2/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 224, 86 GAN fn, tp: 5, 6 GAN f1 score: 0.117 GAN cohens kappa score: 0.059 -> test with 'LR' LR tn, fp: 217, 93 LR fn, tp: 6, 5 LR f1 score: 0.092 LR cohens kappa score: 0.032 LR average precision score: 0.086 -> test with 'GB' GB tn, fp: 286, 24 GB fn, tp: 9, 2 GB f1 score: 0.108 GB cohens kappa score: 0.063 -> test with 'KNN' KNN tn, fp: 215, 95 KNN fn, tp: 7, 4 KNN f1 score: 0.073 KNN cohens kappa score: 0.012 ------ Step 5/5: Slice 3/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 273, 37 GAN fn, tp: 7, 4 GAN f1 score: 0.154 GAN cohens kappa score: 0.106 -> test with 'LR' LR tn, fp: 201, 109 LR fn, tp: 1, 10 LR f1 score: 0.154 LR cohens kappa score: 0.097 LR average precision score: 0.236 -> test with 'GB' GB tn, fp: 288, 22 GB fn, tp: 8, 3 GB f1 score: 0.167 GB cohens kappa score: 0.125 -> test with 'KNN' KNN tn, fp: 253, 57 KNN fn, tp: 6, 5 KNN f1 score: 0.137 KNN cohens kappa score: 0.084 ------ Step 5/5: Slice 4/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1194 synthetic samples -> test with GAN.predict GAN tn, fp: 238, 72 GAN fn, tp: 5, 6 GAN f1 score: 0.135 GAN cohens kappa score: 0.080 -> test with 'LR' LR tn, fp: 206, 104 LR fn, tp: 3, 8 LR f1 score: 0.130 LR cohens kappa score: 0.072 LR average precision score: 0.178 -> test with 'GB' GB tn, fp: 298, 12 GB fn, tp: 10, 1 GB f1 score: 0.083 GB cohens kappa score: 0.048 -> test with 'KNN' KNN tn, fp: 222, 88 KNN fn, tp: 8, 3 KNN f1 score: 0.059 KNN cohens kappa score: -0.002 ------ Step 5/5: Slice 5/5 ------- -> Reset the GAN -> Train generator for synthetic samples -> create 1196 synthetic samples -> test with GAN.predict GAN tn, fp: 253, 53 GAN fn, tp: 6, 3 GAN f1 score: 0.092 GAN cohens kappa score: 0.045 -> test with 'LR' LR tn, fp: 237, 69 LR fn, tp: 3, 6 LR f1 score: 0.143 LR cohens kappa score: 0.097 LR average precision score: 0.139 -> test with 'GB' GB tn, fp: 288, 18 GB fn, tp: 7, 2 GB f1 score: 0.138 GB cohens kappa score: 0.103 -> test with 'KNN' KNN tn, fp: 224, 82 KNN fn, tp: 7, 2 KNN f1 score: 0.043 KNN cohens kappa score: -0.009 ### Exercise is done. -----[ LR ]----- maximum: LR tn, fp: 240, 110 LR fn, tp: 6, 10 LR f1 score: 0.165 LR cohens kappa score: 0.110 LR average precision score: 0.352 average: LR tn, fp: 218.28, 90.92 LR fn, tp: 3.56, 7.04 LR f1 score: 0.129 LR cohens kappa score: 0.074 LR average precision score: 0.160 minimum: LR tn, fp: 200, 69 LR fn, tp: 1, 3 LR f1 score: 0.056 LR cohens kappa score: 0.003 LR average precision score: 0.059 -----[ GB ]----- maximum: GB tn, fp: 301, 32 GB fn, tp: 11, 5 GB f1 score: 0.303 GB cohens kappa score: 0.270 average: GB tn, fp: 291.0, 18.2 GB fn, tp: 8.56, 2.04 GB f1 score: 0.136 GB cohens kappa score: 0.097 minimum: GB tn, fp: 274, 9 GB fn, tp: 6, 0 GB f1 score: 0.000 GB cohens kappa score: -0.041 -----[ KNN ]----- maximum: KNN tn, fp: 253, 101 KNN fn, tp: 10, 7 KNN f1 score: 0.140 KNN cohens kappa score: 0.084 average: KNN tn, fp: 228.04, 81.16 KNN fn, tp: 6.96, 3.64 KNN f1 score: 0.077 KNN cohens kappa score: 0.019 minimum: KNN tn, fp: 209, 57 KNN fn, tp: 4, 1 KNN f1 score: 0.023 KNN cohens kappa score: -0.040 -----[ GAN ]----- maximum: GAN tn, fp: 276, 161 GAN fn, tp: 9, 9 GAN f1 score: 0.160 GAN cohens kappa score: 0.113 average: GAN tn, fp: 228.64, 80.56 GAN fn, tp: 6.2, 4.4 GAN f1 score: 0.095 GAN cohens kappa score: 0.040 minimum: GAN tn, fp: 149, 30 GAN fn, tp: 2, 0 GAN f1 score: 0.000 GAN cohens kappa score: -0.048