Create deploy
Browse files
deploy.py
ADDED
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import pandas as pd
|
2 |
+
from sklearn.model_selection import train_test_split
|
3 |
+
from sklearn.linear_model import LogisticRegression
|
4 |
+
|
5 |
+
datos = pd.read_csv("nueva_base_de_datos.csv", delimiter=',')
|
6 |
+
X = datos.drop('loan_status', axis = 1)
|
7 |
+
y = datos['loan_status']
|
8 |
+
x_train, x_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42, stratify = y)
|
9 |
+
|
10 |
+
# Crea y ajusta el modelo de regresión logística
|
11 |
+
modelo = LogisticRegression()
|
12 |
+
modelo.fit(x_train, y_train)
|
13 |
+
|
14 |
+
import gradio as gr
|
15 |
+
|
16 |
+
def predict(person_income, loan_int_rate, person_age, person_home_ownership_numerica,
|
17 |
+
person_emp_length, loan_intent_numerica, loan_grade_numerica, loan_amnt,
|
18 |
+
cb_person_default_on_file_numerica, cb_person_cred_hist_length):
|
19 |
+
|
20 |
+
print(cb_person_default_on_file_numerica, person_home_ownership_numerica)
|
21 |
+
html = (
|
22 |
+
"<div style='max-width:100%; max-height:360px; overflow:auto'>"
|
23 |
+
+
|
24 |
+
"""<p>Puntajes:</p>
|
25 |
+
<ul>
|
26 |
+
<li>BAJO: 300-579</li>
|
27 |
+
<li>JUSTO: 580-669</li>
|
28 |
+
<li>BUENO: 670-739</li>
|
29 |
+
<li>MUY BUENO: 740-799</li>
|
30 |
+
<li>EXCELENTE: 800-850</li>
|
31 |
+
</ul>"""
|
32 |
+
+ "</div>"
|
33 |
+
)
|
34 |
+
|
35 |
+
df = pd.DataFrame(
|
36 |
+
{
|
37 |
+
'person_age': person_age,
|
38 |
+
'person_income': person_income,
|
39 |
+
'person_emp_length': person_emp_length,
|
40 |
+
'loan_amnt': loan_amnt,
|
41 |
+
'loan_int_rate': loan_int_rate,
|
42 |
+
'cb_person_cred_hist_length': cb_person_cred_hist_length,
|
43 |
+
'person_home_ownership_numerica': person_home_ownership_numerica,
|
44 |
+
'loan_intent_numerica': loan_intent_numerica,
|
45 |
+
'loan_grade_numerica': loan_grade_numerica,
|
46 |
+
'cb_person_default_on_file_numerica': cb_person_default_on_file_numerica,
|
47 |
+
}, index=[0]
|
48 |
+
)
|
49 |
+
|
50 |
+
pred = modelo.predict_proba(df)[0]
|
51 |
+
return 300 + (pred[1] * 550), html
|
52 |
+
|
53 |
+
"""Según lo anterior, las variables categorias quedaron de la siguiente forma numerica.
|
54 |
+
person_home_ownership: -RENT:3 -OWN:2 -MORTAGE:0 -OTHER:1
|
55 |
+
loan_intent: -VENTURE:5 -PERSONAL:4 -EDUCATION:1 -MEDICAL:3 -HOMEIMPROVEMENT:2 -DEBTCONSOLIDATION:0
|
56 |
+
loan_grade: -A:0 -B:1 -C:2 -D:3 -E:4 -F:5 -G:6
|
57 |
+
cb_person_default_on_file: -Y:1 -N:0"""
|
58 |
+
|
59 |
+
inputs = [
|
60 |
+
gr.Slider(1000, 100000, value= 4500, step=500, label='Ingreso Anual'),
|
61 |
+
gr.Slider(0, 25, value= 8.2, label='Tasa de Interes'),
|
62 |
+
gr.Slider(10, 95, value=25, step=1, label='Edad'),
|
63 |
+
gr.Dropdown([('Rentada', 3), ('Propia', 2), ('Hipoteca', 0), ('Otro', 1)], type='index', label='Tipo de Vivienda que posee'),
|
64 |
+
gr.Slider(0, 50, value=6, step=1, label='Años de experiencia laboral'),
|
65 |
+
gr.Dropdown([('Educación', 1), ('Empresa', 5), ('Consolidación de la Deuda', 0), ('Mejora de Vivienda', 2), ('Medico', 3), ('Personal', 4)], type='index', label='Intención del Prestamo'),
|
66 |
+
gr.Dropdown([('A', 0), ('B', 1), ('C', 2), ('D', 3), ('E', 4), ('F', 5), ('G', 6)], type='index', label='Grado del Prestamo'),
|
67 |
+
gr.Slider(1000, 100000, value= 4500, step=500, label='Monto del Prestamo'),
|
68 |
+
gr.Dropdown([('Si', 0), ('No', 1), ('No', 1)], type='index', label='Hay incumplimientos en el historial crediticio ?'),
|
69 |
+
gr.Slider(0, 35, value=4, step=1, label='Duración del Historial Crediticio'),
|
70 |
+
]
|
71 |
+
|
72 |
+
demo = gr.Interface(fn=predict, inputs=inputs, outputs=["label", "html"], title='Modelo de Riesgo: ScoreCard')
|
73 |
+
demo.launch(share=True, debug=True)
|
74 |
+
|