Revisiting Warfarin Dosing Using Machine Learning Techniques

Table 1

Dataset description.

Continuous variables
Target international normalized ratio	Mean		2.5
	Std. deviation		0.1
	Minimum		1.8
	Maximum		3.5
Body surface area	Mean		1.94
	Std. deviation		0.3
	Minimum		1.2
	Maximum		3.4
Categorical variables
	Values	Frequency	Percent
Gender	0	1822	43.00%
	1	2415	57.00%
Race	1	2663	62.85%
	2	656	15.48%
	3	918	21.67%
Deep vein thrombosis and pulmonary embolism	0	3846	90.77%
	1	391	9.23%
Diabetes	0	3500	82.61%
	1	737	17.39%
Congestive heart failure	0	3492	82.42%
	1	745	17.58%
Valve replacement	0	3243	76.54%
	1	994	23.46%
Aspirin	0	3199	75.50%
	1	1038	24.50%
Simvastatin	0	3608	85.15%
	1	629	14.85%
Atorvastatin	0	3810	89.92%
	1	427	10.08%
Fluvastatin	0	4220	99.60%
	1	17	0.40%
Lovastatin	0	4153	98.02%
	1	84	1.98%
Pravastatin	0	4121	97.26%
	1	116	2.74%
Rosuvastatin	0	4208	99.32%
	1	29	0.68%
Amiodarone	0	3984	94.03%
	1	253	5.97%
Carbamazepine	0	4195	99.01%
	1	42	0.99%
Phenytoin	0	4197	99.06%
	1	40	0.94%
Rifampin	0	4231	99.86%
	1	6	0.14%
Sulfonamide Antibiotics	0	4214	99.46%
	1	23	0.54%
Macrolide antibiotics	0	4225	99.72%
	1	12	0.28%
Antifungal azoles	0	4210	99.36%
	1	27	0.64%
Smoker	0	3733	88.10%
	1	504	11.90%
Enzyme	0	4150	97.95%
	1	87	2.05%
Patient class	0	2111	49.82%
	1	2126	50.18%
Age	1	9	0.21%
	2	94	2.22%
	3	189	4.46%
	4	444	10.48%
	5	806	19.02%
	6	1023	24.14%
	7	1133	26.74%
	8	511	12.06%
	9	28	0.66%