Hematoma expansion prediction based on SMOTE and XGBoost algorithm

Table 2 Variable description

Characteristics	Case group	Control group	P value
Age	59.28±13.51	60.12±12.41	0.03
Gender (Male)	53(46.5%)	221(49.5%)	0.58
Diabetes mellitus	9(7.9%)	64(15.1%)	0.19
Alcohol use	12(10.5%)	43(9.4%)	0.19
Admission SBP	173.63±20.57	168.63±19.24	0.02
Admission DBP	101.25±14.39	98.56±12.31	0.04
ICH position	79(46.8%)	196(41.8%)	0.56
Left/Right	62(47.7%)	182(38.9%)	0.45
Shape score	2.81±1.63	2.80±1.61	0.34
Heterogeneity	3.72±1.68	3.64±1.59	0.43
Island sign	45(39.5%)	112(23.9%)	0.28
Satellite sign	10(8.8%)	45(9.6%)	0.24
Black hole sign	15(13.2%)	47(15.2%)	0.17
Blend sign	60(52.6%)	192(46.8%)	0.12
Swirl sign	13(11.4%)	52(12.8%)	0.61
IVH	39(34.2%)	107(39.7%)	0.54
SAH	7(6.1%)	22(7.2%)	0.77
MLS	34(29.8%)	102(28.8%)	0.89
Admission GCS score	10.49±2.99	10.27±2.21	0.19
Hematoma volume	15.73±15.26	14.78±12.37	0.01

The first column corresponds to the name of the indicators used, and the second and third columns correspond to the statistics of each indicator in the case group and the control group, respectively. The statistic of numerical indicators such as age is expressed by means variance, and the statistic of binary indicators such as alcohol use is expressed by number and ratio. P values in the last column are used to determine whether various indicators are significant

ISSN: 1472-6947