Machine learning-based nomogram for predicting depressive symptoms in women: A cross-sectional study in Guangdong Province, China

Table 1 Demographics and baseline characteristics, n (%)

Characteristic	Overall, N = 7609	No depressive symptoms, n = 7090	Depressive symptoms, n = 519	Statistic	P value1
Residence				0.48	0.488
Urban area	5933 (78.0)	5522 (77.9)	411 (79.2)
Rural area	1676 (22.0)	1568 (22.1)	108 (20.8)
Age, years				47.52	< 0.001
18-39	5552 (73.0)	5106 (72.0)	446 (85.9)
40-59	1901 (25.0)	1833 (25.9)	68 (13.1)
60-85	156 (2.1)	151 (2.1)	5 (1.0)
Education, years				2.78	0.250
≤ 9	815 (10.7)	769 (10.8)	46 (8.9)
10-15	4094 (53.8)	3818 (53.9)	276 (53.2)
≥ 16	2700 (35.5)	2503 (35.3)	197 (38.0)
Marital				67.87	< 0.001
Single/unmarried	2431 (31.9)	2186 (30.8)	245 (47.2)
Married	4973 (65.4)	4720 (66.6)	253 (48.7)
Widowed/divorced/separated	205 (2.7)	184 (2.6)	21 (4.0)
Income class, yuan				1.23	0.541
< 3500	2935 (38.6)	2723 (38.4)	212 (40.8)
3500-6000	3120 (41.0)	2916 (41.1)	204 (39.3)
> 6000	1554 (20.4)	1451 (20.5)	103 (19.8)
Drink, time/week				59.46	< 0.00
< 1	7082 (93.1)	6642 (93.7)	440 (84.8)
≥ 1	527 (6.9)	448 (6.3)	79 (15.2)
Tea, times/week				0.32	0.569
< 4	4824 (63.4)	4501 (63.5)	323 (62.2)
≥ 4	2785 (36.6)	2589 (36.5)	196 (37.8)
Exercise				76.01	< 0.001
Hardly exercise	2347 (30.8)	2103 (29.7)	244 (47.0)
1-3 times/month	2452 (32.2)	2308 (32.6)	144 (27.7)
1-2 times/week	1426 (18.7)	1345 (19.0)	81 (15.6)
≥ 3 times/week	1384 (18.2)	1334 (18.8)	50 (9.6)
Chronic disease				244.42	< 0.001
Yes	5745 (75.5)	5501 (77.6)	244 (47.0)
No	1864 (24.5)	1589 (22.4)	275 (53.0)
BMI				14.00	0.003
18.5-23.9	4964 (65.2)	4636 (65.4)	328 (63.2)
< 18.5	1329 (17.5)	1212 (17.1)	117 (22.5)
24.0-27.9	1061 (13.9)	1007 (14.2)	54 (10.4)
> 28.0	255 (3.4)	235 (3.3)	20 (3.9)
Anxiety symptoms				2009.91	< 0.001
No	7263 (95.5)	6973 (98.3)	290 (55.9)
Yes	346 (4.5)	117 (1.7)	229 (44.1)
Sleep latency, minutes				222.56	< 0.001
≤ 15	2744 (36.1)	2647 (37.3)	97 (18.7)
16-30	3687 (48.5)	3457 (48.8)	230 (44.3)
31-60	733 (9.6)	625 (8.8)	108 (20.8)
> 60	445 (5.8)	361 (5.1)	84 (16.2)
Sleep duration, hours				216.64	< 0.001
≥ 7	3014 (39.6)	2895 (40.8)	119 (22.9)
6-6.99	3980 (52.3)	3705 (52.3)	275 (53.0)
5-5.99	461 (6.1)	370 (5.2)	91 (17.5)
< 5	154 (2.0)	120 (1.7)	34 (6.6)
Sleep efficiency, %				95.87	< 0.001
> 85	5948 (78.2)	5606 (79.1)	342 (65.9)
75-84	1119 (14.7)	1030 (14.5)	89 (17.1)
65-74	319 (4.2)	275 (3.9)	44 (8.5)
< 65	223 (2.9)	179 (2.5)	44 (8.5)
Smoke				92.97	< 0.001
No	7481 (98.3)	6998 (98.7)	483 (93.1)
Yes	128 (1.7)	92 (1.3)	36 (6.9)
Insomnia				1135.26	< 0.001
No	6179 (81.2)	6047 (85.3)	132 (25.4)
Yes	1430 (18.8)	1043 (14.7)	387 (74.6)

¹Pearson’s χ² test.

BMI: Body mass index.

Table 2 The performance metrics of three machine learning models

	Accuracy	Precision	Recall	F1 score	AUC
XGBoost	0.941	0.603	0.429	0.501	0.862
SVM	0.946	0.707	0.371	0.487	0.849
LightGBM	0.944	0.670	0.378	0.483	0.867

XGBoost: Extreme gradient boosting; SVM: Support vector machine; LightGBM: Light gradient boosting machine; AUC: Area under the curve.