Computational and Mathematical Methods in Medicine / 2016 / Article / Tab 1 / Research Article
Survival Prediction and Feature Selection in Patients with Breast Cancer Using Support Vector Regression Table 1 Comparison of characteristics of dead and alive patients in the BC dataset, and the relation between the features and the estimated prognostic based on standard linear SVR.
Independent variables Category Characteristics of patients in dataset Estimated prognostic index in SVR model Alive Dead value Count Mean (SD) Count Mean (SD) Pearson correlation (with prognostic index) Age of diagnosis 341 45.33 (10.26) 201 47.30 (11.64) −0.22 0.001 Count Percent Count Percent Mean prognostic index SD prognostic index Tumor size (cm) 2230 70.12 98 29.88 −15.98 12.18 0.313 2–5 103 56.28 80 43.72 −16.47 12.59 >5 8 25.80 23 74.2 −23.18 8.12 Involved lymph nodes ( ) <4 198 69.72 86 30.28 −13.44 11.87 <0.001 4103 62.05 63 37.95 −21.14 11.32 Unknown 40 43.48 52 56.52 Marital status Married 326 62.69 194 37.31 −16.20 12.21 0.354 Single 15 68.18 7 31.82 −20.29 12.93 Family history Yes 300 62.11 183 37.89 −16.46 7.12 0.418 No 41 69.49 18 30.51 −14.94 12.58 Type of surgery Lumpectomy 18 64.24 10 35.76 −13.34 9.19 0.031 Radical mastectomy 250 61.27 158 38.73 −15.47 12.13 Segmental mastectomy 26 63.41 15 36.59 −18.64 15.79 simple mastectomy 47 72.30 18 27.70 −22.62 9.75 Histological type Ductal 301 62.57 180 37.43 −17.18 12.26 0.022 Lobular 21 61.76 13 38.24 −10.63 10.69 Medullar 19 70.37 8 29.63 −10.76 10.81 Metastases status Present 51 36.42 89 63.58 −34.32 10.82 <0.001 Absent 290 72.13 112 27.87 −12.83 8.99 HER2 Positive 110 65.47 58 34.53 −20.89 11.62 <0.001 Negative 84 60.00 56 40.00 −10.02 10.09 Unknown 147 62.82 87 37.18 ER status Positive 106 63.85 60 36.15 −17.54 12.35 0.148 Negative 126 68.10 59 31.90 −15.24 12.06 Unknown 109 57.06 82 42.94 PR status Positive 85 59.44 58 40.56 −20.39 11.57 <0.001 Negative 142 71.71 56 28.29 −13.37 11.88 Unknown 114 56.71 87 43.29
value < 0.05, value < 0.01, value < 0.001.