Prediksi Stroke Berbasis Decision Tree

Analisis data kesehatan komprehensif menggunakan dataset healthcare-dataset-stroke-data.csv. Algoritma Decision Tree dikonfigurasi dengan teknik pembebanan kelas seimbang (Class Weight Balanced) untuk menangani ketidakseimbangan kelas imbalanced yang ekstrem pada variabel target Stroke

Jumlah data

5.110

11 kolom

Kasus stroke

249

4.87% dari total

Rata-rata umur

43.2

penderita stroke: 67.7

Rata-rata gula darah

106.1

penderita stroke: 132.5

Performa Model (data uji)

Akurasi

72.5%

Recall (stroke)

72.0%

Precision

11.9%

F1-score

20.4%

ROC-AUC

0.783

Data latih 4088 · data uji 1022 (split 80/20, stratified). Karena kasus stroke hanya ~5%, model memakai class_weight = balanced agar lebih peka mendeteksi stroke (recall tinggi).

Confusion Matrix

Prediksi: Tidak

Prediksi: Stroke

Aktual: Tidak

705

267

Aktual: Stroke

Model berhasil mendeteksi 36 dari 50 kasus stroke pada data uji.

Kepentingan Fitur

Umur

77.4%

BMI

11.3%

Gula darah

5.9%

Penyakit jantung

2.1%

Status merokok

1.6%

Jenis pekerjaan

1.5%

Tempat tinggal

0.1%

Hipertensi

0.1%

Jenis kelamin

0.0%

Status menikah

0.0%

Statistik Deskriptif

Statistik	Umur	Gula darah	BMI
count	5110.0	5110.0	4909.0
mean	43.23	106.15	28.89
std	22.61	45.28	7.85
min	0.08	55.12	10.3
25%	25.0	77.24	23.5
50%	45.0	91.88	28.1
75%	61.0	114.09	33.1
max	82.0	271.74	97.6

201 data BMI kosong (N/A) diisi dengan nilai median saat pelatihan model.

Aturan Decision Tree (4 level teratas)

|--- num__age <= 47.5
|   |--- num__age <= 37.5
|   |   |--- num__age <= 1.4
|   |   |   |--- num__bmi <= 19.8
|   |   |   |   |--- class: 0
|   |   |   |--- num__bmi >  19.8
|   |   |   |   |--- class: 1
|   |   |--- num__age >  1.4
|   |   |   |--- class: 0
|   |--- num__age >  37.5
|   |   |--- num__bmi <= 45.7
|   |   |   |--- num__avg_glucose_level <= 59.0
|   |   |   |   |--- class: 0
|   |   |   |--- num__avg_glucose_level >  59.0
|   |   |   |   |--- cat__smoking_status_formerly smoked <= 0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- cat__smoking_status_formerly smoked >  0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |--- num__bmi >  45.7
|   |   |   |--- class: 1
|--- num__age >  47.5
|   |--- num__age <= 67.5
|   |   |--- num__bmi <= 27.2
|   |   |   |--- num__bmi <= 26.4
|   |   |   |   |--- cat__smoking_status_smokes <= 0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- cat__smoking_status_smokes >  0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |--- num__bmi >  26.4
|   |   |   |   |--- class: 0
|   |   |--- num__bmi >  27.2
|   |   |   |--- num__bmi <= 28.6
|   |   |   |   |--- num__age <= 53.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- num__age >  53.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |--- num__bmi >  28.6
|   |   |   |   |--- num__heart_disease <= 0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- num__heart_disease >  0.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |--- num__age >  67.5
|   |   |--- num__age <= 73.5
|   |   |   |--- num__avg_glucose_level <= 177.0
|   |   |   |   |--- num__avg_glucose_level <= 94.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- num__avg_glucose_level >  94.5
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |--- num__avg_glucose_level >  177.0
|   |   |   |   |--- num__bmi <= 28.1
|   |   |   |   |   |--- class: 1
|   |   |   |   |--- num__bmi >  28.1
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |--- num__age >  73.5
|   |   |   |--- num__bmi <= 34.2
|   |   |   |   |--- num__bmi <= 27.9
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |   |--- num__bmi >  27.9
|   |   |   |   |   |--- truncated branch of depth 2
|   |   |   |--- num__bmi >  34.2
|   |   |   |   |--- cat__work_type_Private <= 0.5
|   |   |   |   |   |--- class: 0
|   |   |   |   |--- cat__work_type_Private >  0.5
|   |   |   |   |   |--- class: 1

Cuplikan Data (10 baris pertama)

gender	age	hypertension	heart_disease	ever_married	work_type	Residence_type	avg_glucose_level	bmi	smoking_status	stroke
Male	67.0	0	1	Yes	Private	Urban	228.69	36.6	formerly smoked	1
Female	61.0	0	0	Yes	Self-employed	Rural	202.21	N/A	never smoked	1
Male	80.0	0	1	Yes	Private	Rural	105.92	32.5	never smoked	1
Female	49.0	0	0	Yes	Private	Urban	171.23	34.4	smokes	1
Female	79.0	1	0	Yes	Self-employed	Rural	174.12	24.0	never smoked	1
Male	81.0	0	0	Yes	Private	Urban	186.21	29.0	formerly smoked	1
Male	74.0	1	1	Yes	Private	Rural	70.09	27.4	never smoked	1
Female	69.0	0	0	No	Private	Urban	94.39	22.8	never smoked	1
Female	59.0	0	0	Yes	Private	Rural	76.15	N/A	Unknown	1
Female	78.0	0	0	Yes	Private	Urban	58.57	24.2	Unknown	1