Spaces:
Sleeping
Sleeping
scorpion237
commited on
Commit
•
6a19a00
1
Parent(s):
74fa9ef
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,441 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st, base64
|
2 |
+
import pandas as pd, seaborn as sns
|
3 |
+
import os, matplotlib.pyplot as plt
|
4 |
+
import pickle, numpy as np
|
5 |
+
from sklearn.preprocessing import StandardScaler
|
6 |
+
from sklearn.metrics import classification_report, accuracy_score, confusion_matrix
|
7 |
+
|
8 |
+
|
9 |
+
# image de fond
|
10 |
+
def add_bg_from_local(image_file):
|
11 |
+
with open(image_file, "rb") as image_file:
|
12 |
+
encoded_string = base64.b64encode(image_file.read())
|
13 |
+
st.markdown(
|
14 |
+
f"""
|
15 |
+
<style>
|
16 |
+
.stApp {{
|
17 |
+
background-image: url(data:image/{"png"};base64,{encoded_string.decode()});
|
18 |
+
background-size: cover
|
19 |
+
}}
|
20 |
+
</style>
|
21 |
+
""",
|
22 |
+
unsafe_allow_html=True
|
23 |
+
)
|
24 |
+
add_bg_from_local('./images/route.png')
|
25 |
+
|
26 |
+
|
27 |
+
fig = plt.figure(figsize=(10, 10))
|
28 |
+
_, middle, _ = st.columns((2, 3, 2))
|
29 |
+
with middle:
|
30 |
+
st.title(":orange[_Scoring App_]")
|
31 |
+
# path du dossier data
|
32 |
+
path = ".\data"
|
33 |
+
|
34 |
+
# fonction pour loader le dataset
|
35 |
+
@st.cache_data
|
36 |
+
def load_data(file_path):
|
37 |
+
return pd.read_csv(os.path.join(path, file_path))
|
38 |
+
|
39 |
+
# convertir dataframe en csv
|
40 |
+
def convert_df_to_csv(frame):
|
41 |
+
return frame.to_csv(index=False).encode("utf-8")
|
42 |
+
|
43 |
+
# fonction principale
|
44 |
+
st.sidebar.image(r"./images/picture1.png")
|
45 |
+
def main():
|
46 |
+
st.markdown("<h2 style = 'text-align:center; \
|
47 |
+
color:green;'> Classification pour l'octroi de credit </h2>", unsafe_allow_html = True)
|
48 |
+
|
49 |
+
# charger le fichier
|
50 |
+
uploaded_file = st.sidebar.file_uploader("Upload your input CSV file", type=["csv"])
|
51 |
+
|
52 |
+
# creation du menu
|
53 |
+
menu = ["Home", "Data Exploration", "Data Visualisation", "Make prediction"]
|
54 |
+
choice = st.sidebar.selectbox("Select menu", menu)
|
55 |
+
|
56 |
+
# charger le jeu de donnees
|
57 |
+
data = load_data("loan.csv")
|
58 |
+
|
59 |
+
# supprime la colonne Loan_ID
|
60 |
+
data.drop("Loan_ID", axis=1, inplace=True)
|
61 |
+
if choice == "Home":
|
62 |
+
st.write("Nous avons develeopper pour ce projet un model de classification\
|
63 |
+
qui permet, sur la base de certaines variables, de determiner si oui ou non\
|
64 |
+
il est envisageable d'octroyer un pret bancaire a une tierce personne.")
|
65 |
+
|
66 |
+
st.subheader(":orange[__Presentation du jeu de donnee__] :memo:")
|
67 |
+
|
68 |
+
st.markdown("Le jeu de donnees comporte 614 lignes et 13 colonnes. **Loan_Status**\
|
69 |
+
est la variables a predire (categorielle a deuc classe: **Y** pour le pret a ete \
|
70 |
+
octroyer et **N**) pour le contraire. afin d'avoir les reultats les plus\
|
71 |
+
optimaux possibles, nous allons dans un premier temps faire une \
|
72 |
+
***Analyse exploratoire** de nos donnees. Par suite nous passerons\
|
73 |
+
a la phase de preparation des donnees pour afin finir avec \
|
74 |
+
la phase de creation et optimisation des models.\
|
75 |
+
`Si vous televerser un fichier au format csv, vous avez la\
|
76 |
+
possibilite de comparer les prediction pour chaque\
|
77 |
+
model et de telechager le fichier csv correspondant.`")
|
78 |
+
#st.image("./images/processor.jpg")
|
79 |
+
|
80 |
+
if choice == "Data Exploration":
|
81 |
+
st.subheader(":orange[_Data Exploration_] :bar_chart:")
|
82 |
+
# afficher les donnees
|
83 |
+
st.write(data.head())
|
84 |
+
|
85 |
+
# valeurs manquante
|
86 |
+
if st.sidebar.checkbox("Valeur Manquante"):
|
87 |
+
st.subheader(":orange[Valeur Manquante]")
|
88 |
+
na_count = data.isnull().sum().to_frame(name='count')
|
89 |
+
na_per = (data.isnull().sum().to_frame(name='percentage %')/data.shape[0]*100).round(2)
|
90 |
+
st.write(pd.concat([na_count, na_per], axis=1).sort_values(by='count', ascending=False).T)
|
91 |
+
|
92 |
+
# valeur unique par colonne
|
93 |
+
if st.sidebar.checkbox("Valeur Unique par colonnes"):
|
94 |
+
st.subheader(":orange[Valeur Unique par colonnes]")
|
95 |
+
only = data.nunique().sort_values(ascending=False).to_frame(name='count')
|
96 |
+
perc = (data.nunique().sort_values(ascending=False).to_frame(name='percentage %')/data.shape[0]*100).round(2)
|
97 |
+
dtype = data.dtypes.to_frame(name='dtypes')
|
98 |
+
st.write(pd.concat([only, perc, dtype], axis=1).T)
|
99 |
+
|
100 |
+
# statistique sommaire
|
101 |
+
if st.sidebar.checkbox("Statistiques somaire"):
|
102 |
+
st.subheader(":orange[Statistiques sommaire]")
|
103 |
+
st.write(data.describe())
|
104 |
+
|
105 |
+
# matrice de correlation
|
106 |
+
if st.sidebar.checkbox("Matrice de correlation"):
|
107 |
+
fig = plt.figure(figsize=(7,5))
|
108 |
+
st.subheader(":orange[Matrice de correlation]")
|
109 |
+
st.write(sns.heatmap(data.corr(), annot=True, vmin=-1, vmax=1, cmap='ocean'))
|
110 |
+
st.pyplot(fig)
|
111 |
+
plt.show()
|
112 |
+
|
113 |
+
if choice == "Data Visualisation":
|
114 |
+
st.subheader(":orange[_Data Visualisation_] :chart:")
|
115 |
+
if st.sidebar.checkbox("Analyse Univariee"):
|
116 |
+
# selection des variables qualitatives
|
117 |
+
categorical_columns = data.select_dtypes(include='object').columns.tolist()
|
118 |
+
st.write("Liste des variables qaulitatives")
|
119 |
+
st.write(categorical_columns)
|
120 |
+
fig = plt.figure(figsize=(14, 8))
|
121 |
+
sns.set_theme(context='notebook', style='darkgrid', palette='deep', font='sans-serif', font_scale=1, color_codes=True, rc=None)
|
122 |
+
for idx, col in enumerate(categorical_columns[:-1]):
|
123 |
+
plt.subplot(2, 3, idx+1)
|
124 |
+
sns.countplot(data=data, x=col, hue="Loan_Status")
|
125 |
+
sns.countplot(data=data, x='Loan_Status')
|
126 |
+
st.pyplot(fig)
|
127 |
+
plt.show()
|
128 |
+
|
129 |
+
# selection des variables quantitatives
|
130 |
+
numerical_columns = data.select_dtypes(include='number').columns.tolist()
|
131 |
+
st.write("Liste des variables quantitatives")
|
132 |
+
st.write(numerical_columns)
|
133 |
+
fig = plt.figure(figsize=(15,7))
|
134 |
+
for idx, col in enumerate(numerical_columns):
|
135 |
+
plt.subplot(2,3, idx+1)
|
136 |
+
plt.hist(data[col], density=True)
|
137 |
+
sns.kdeplot(data=data, x=col)
|
138 |
+
plt.title(col)
|
139 |
+
#plt.subplots_adjust(hspace=0.5)
|
140 |
+
plt.tight_layout(h_pad=2, w_pad=3., rect=(1,1,2,2))
|
141 |
+
st.pyplot(fig)
|
142 |
+
plt.show()
|
143 |
+
|
144 |
+
if st.sidebar.checkbox("Analyse bivariee"):
|
145 |
+
st.subheader(":orange[Analyse bivariee]")
|
146 |
+
numerical_columns = data.select_dtypes(include='number').columns.tolist()
|
147 |
+
fig = plt.figure(figsize = (14, 8))
|
148 |
+
for idx, num_col in enumerate(numerical_columns[:-2]):
|
149 |
+
plt.subplot(2, 2, idx+1)
|
150 |
+
sns.boxplot(y=num_col, data=data, x='Loan_Status')
|
151 |
+
plt.tight_layout(h_pad=2, w_pad=3., rect=(1,1,2,2))
|
152 |
+
st.pyplot(fig)
|
153 |
+
plt.show()
|
154 |
+
|
155 |
+
if choice == "Make prediction":
|
156 |
+
st.subheader(":orange[Make prediction] :fleur_de_lis:")
|
157 |
+
if uploaded_file is not None:
|
158 |
+
data = pd.read_csv(uploaded_file)
|
159 |
+
|
160 |
+
# data preprocessing
|
161 |
+
from sklearn.impute import SimpleImputer
|
162 |
+
try:
|
163 |
+
data.drop(["Loan_ID"], axis=1, inplace=True)
|
164 |
+
except:
|
165 |
+
pass
|
166 |
+
# encodage
|
167 |
+
data_encoded = pd.get_dummies(data, drop_first=True)
|
168 |
+
st.subheader(":orange[Donnees encodees]")
|
169 |
+
st.write(data_encoded)
|
170 |
+
|
171 |
+
# separation du jeu de donnee
|
172 |
+
X, y = data_encoded.drop(["Loan_Status_Y"], axis=1), data_encoded["Loan_Status_Y"]
|
173 |
+
|
174 |
+
# traintement des valeurs manquantes
|
175 |
+
sp = SimpleImputer(strategy="most_frequent")
|
176 |
+
X = sp.fit_transform(X)
|
177 |
+
|
178 |
+
# mis a l'echelle des variables
|
179 |
+
std = StandardScaler()
|
180 |
+
X = std.fit_transform(X)
|
181 |
+
|
182 |
+
# Prediction
|
183 |
+
# Random Forest predictor
|
184 |
+
if st.sidebar.checkbox("Random Forest"):
|
185 |
+
st.subheader(":orange[Random Forest] :sunglasses:")
|
186 |
+
rf = pickle.load(open("scoring_rf.pkl", "rb"))
|
187 |
+
pred = rf.predict(X)
|
188 |
+
pred_proba = rf.predict_proba(X)
|
189 |
+
st.subheader(':green[Prediction]')
|
190 |
+
loan_status = np.array(['N','Y'])
|
191 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
192 |
+
df = pd.concat([data, prediction], axis=1)
|
193 |
+
st.write(df)
|
194 |
+
# download frame
|
195 |
+
csv = convert_df_to_csv(df)
|
196 |
+
st.download_button("Press to Download",
|
197 |
+
csv,
|
198 |
+
"random_forest.csv",
|
199 |
+
"text/csv",
|
200 |
+
key='download_csv')
|
201 |
+
|
202 |
+
st.text("Model report : \n " + classification_report(y, pred))
|
203 |
+
|
204 |
+
# Accuracy score
|
205 |
+
rf_score = accuracy_score(pred,y)
|
206 |
+
st.write(":green[score d'exactitude]")
|
207 |
+
st.write(f"{round(rf_score*100,2)}% d'exactitude")
|
208 |
+
st.subheader(':green[Prediction Probability]')
|
209 |
+
st.write(pred_proba)
|
210 |
+
|
211 |
+
# Linear Discriminant Analysis
|
212 |
+
if st.sidebar.checkbox("Discriminant Analysis"):
|
213 |
+
st.subheader(":orange[Discriminant Analysis] :sunglasses:")
|
214 |
+
lda = pickle.load(open("scoring_lda.pkl", "rb"))
|
215 |
+
pred = lda.predict(X)
|
216 |
+
pred_proba = lda.predict_proba(X)
|
217 |
+
st.subheader(':green[Prediction]')
|
218 |
+
loan_status = np.array(['N','Y'])
|
219 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
220 |
+
df = pd.concat([data, prediction], axis=1)
|
221 |
+
st.write(df)
|
222 |
+
# download
|
223 |
+
csv = convert_df_to_csv(df)
|
224 |
+
st.download_button("Press to Download",
|
225 |
+
csv,
|
226 |
+
"discriminant.csv",
|
227 |
+
"text/csv",
|
228 |
+
key='download_csv')
|
229 |
+
st.text("Model report : \n " + classification_report(y, pred))
|
230 |
+
|
231 |
+
# Accuracy score
|
232 |
+
lda_score = accuracy_score(pred,y)
|
233 |
+
st.subheader(":green[score d'exactitude]")
|
234 |
+
st.write(f"{round(lda_score*100,2)}% d'exactitude")
|
235 |
+
st.subheader(':green[Prediction Probability]')
|
236 |
+
st.write(pred_proba)
|
237 |
+
|
238 |
+
# matrice de confusion
|
239 |
+
fig = plt.figure(figsize=(2,1))
|
240 |
+
cm = confusion_matrix(y, pred)
|
241 |
+
st.subheader(":green[Matrice de confusion]")
|
242 |
+
sns.heatmap(cm, annot=True, cmap='Dark2')
|
243 |
+
st.pyplot(fig)
|
244 |
+
plt.plot()
|
245 |
+
|
246 |
+
# XGBoost
|
247 |
+
if st.sidebar.checkbox("XGBoost"):
|
248 |
+
st.subheader(":orange[XGBoost] :sunglasses:")
|
249 |
+
xg = pickle.load(open("scoring_xg.pkl", "rb"))
|
250 |
+
pred = xg.predict(X)
|
251 |
+
pred_proba = xg.predict_proba(X)
|
252 |
+
st.subheader(':green[Prediction]')
|
253 |
+
loan_status = np.array(['N','Y'])
|
254 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
255 |
+
df = pd.concat([data, prediction], axis=1)
|
256 |
+
st.write(df)
|
257 |
+
# download
|
258 |
+
csv = convert_df_to_csv(df)
|
259 |
+
st.download_button("Press to Download",
|
260 |
+
csv,
|
261 |
+
"xgboost.csv",
|
262 |
+
"text/csv",
|
263 |
+
key='download_csv')
|
264 |
+
st.text("Model report : \n " + classification_report(y, pred))
|
265 |
+
|
266 |
+
# Accuracy score
|
267 |
+
xg_score = accuracy_score(pred,y)
|
268 |
+
st.subheader(":green[score d'exactitude]")
|
269 |
+
st.write(f"{round(xg_score*100,2)}% d'exactitude")
|
270 |
+
st.subheader(':green[Prediction Probability]')
|
271 |
+
st.write(pred_proba)
|
272 |
+
|
273 |
+
# ANN
|
274 |
+
if st.sidebar.checkbox("Neural Network"):
|
275 |
+
st.subheader(":orange[Neural Network] :sunglasses:")
|
276 |
+
ann = pickle.load(open("scoring_ann.pkl", "rb"))
|
277 |
+
pred = ann.predict(X)
|
278 |
+
pred_proba = ann.predict_proba(X)
|
279 |
+
st.subheader(':green[Prediction]')
|
280 |
+
loan_status = np.array(['N','Y'])
|
281 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
282 |
+
df = pd.concat([data, prediction], axis=1)
|
283 |
+
st.write(df)
|
284 |
+
# download
|
285 |
+
csv = convert_df_to_csv(df)
|
286 |
+
st.download_button("Press to Download",
|
287 |
+
csv,
|
288 |
+
"neural_network.csv",
|
289 |
+
"text/csv",
|
290 |
+
key='download_csv')
|
291 |
+
st.text("Model report : \n " + classification_report(y, pred))
|
292 |
+
|
293 |
+
# Accuracy score
|
294 |
+
ann_score = accuracy_score(pred,y)
|
295 |
+
st.subheader(":green[score d'exactitude]")
|
296 |
+
st.write(f"{round(ann_score*100,2)}% d'exactitude")
|
297 |
+
st.subheader(':green[Prediction Probability]')
|
298 |
+
st.write(pred_proba)
|
299 |
+
|
300 |
+
else:
|
301 |
+
def user_input_features():
|
302 |
+
gender = st.sidebar.selectbox('Gender',('Male','Female'))
|
303 |
+
married = st.sidebar.selectbox('Married',('Yes','No'))
|
304 |
+
depedents = st.sidebar.selectbox('Dependent',(0, 1, 2, "3+"))
|
305 |
+
education = st.sidebar.selectbox('Education',('Graduate','Not Graduate'))
|
306 |
+
self_employed = st.sidebar.selectbox('Self_employed',('Yes','No'))
|
307 |
+
applicanincome = st.sidebar.slider('ApplicanIncome', 150, 81000)
|
308 |
+
coapplicanincome = st.sidebar.slider('CoapplicanIncome', 0, 42000)
|
309 |
+
loan_amount = st.sidebar.slider('LoanAmount', 0, 800)
|
310 |
+
loan_amount_term = st.sidebar.slider('Loan_Amount_Term', 10, 500)
|
311 |
+
credit_history = st.sidebar.selectbox('Credi_History', (0, 1))
|
312 |
+
property_area = st.sidebar.selectbox('Property_Area', ("Urban", "Rural", "Semiurban"))
|
313 |
+
|
314 |
+
if gender == "Male":
|
315 |
+
gender = 1
|
316 |
+
else:
|
317 |
+
gender = 0
|
318 |
+
|
319 |
+
if married == 'Yes':
|
320 |
+
married = 1
|
321 |
+
else:
|
322 |
+
married = 0
|
323 |
+
|
324 |
+
depedents_1, depedents_2, depedents_3 = 0,0,0
|
325 |
+
if depedents == 1:
|
326 |
+
depedents_1=1
|
327 |
+
elif depedents == 2:
|
328 |
+
depedents_2=1
|
329 |
+
elif depedents > 2 :
|
330 |
+
depedents_3=1
|
331 |
+
|
332 |
+
if education == "Not Graduate":
|
333 |
+
education=1
|
334 |
+
else:
|
335 |
+
education=0
|
336 |
+
|
337 |
+
if self_employed == "Yes":
|
338 |
+
self_employed = 1
|
339 |
+
else:
|
340 |
+
self_employed = 0
|
341 |
+
|
342 |
+
property_urban, property_semiurban = 0, 0
|
343 |
+
if property_area == "Semiurban":
|
344 |
+
property_semiurban = 1
|
345 |
+
elif property_area == "Urban":
|
346 |
+
property_urban == 1
|
347 |
+
|
348 |
+
data = { 'ApplicationIncome': (applicanincome - 5403)/6109,
|
349 |
+
'CoapplicationIncome': (coapplicanincome - 1621) / 2926,
|
350 |
+
'LoanAmount': (loan_amount -146)/85,
|
351 |
+
'Loan_Amount_Term': (loan_amount_term - 342)/65,
|
352 |
+
'Credi_History': (credit_history -0.84)/0.35,
|
353 |
+
'Gender_Male': gender,
|
354 |
+
'Married_Yes': married,
|
355 |
+
'Depedents_1': depedents_1,
|
356 |
+
'Depedents_2': depedents_2,
|
357 |
+
'Depedents_3+': depedents_3,
|
358 |
+
'Education_Not_Graduate': education,
|
359 |
+
'Self_Employed_Yes': self_employed,
|
360 |
+
'Property_Area_Semiurban': property_semiurban,
|
361 |
+
'Property_Area_Urban': property_urban
|
362 |
+
}
|
363 |
+
features = pd.DataFrame(data, index=[0])
|
364 |
+
return features
|
365 |
+
data_input = user_input_features()
|
366 |
+
|
367 |
+
# Random Forest
|
368 |
+
if st.sidebar.checkbox("Random Forest"):
|
369 |
+
st.subheader(":orange[Random Forest]")
|
370 |
+
rf = pickle.load(open("scoring_rf.pkl", "rb"))
|
371 |
+
pred = rf.predict(data_input)
|
372 |
+
if pred == 1:
|
373 |
+
st.write(":orange[__Le pret peut etre octroyer__] :white_check_mark:")
|
374 |
+
else:
|
375 |
+
st.write(":red[__Desole,...__] :disappointed:")
|
376 |
+
pred_proba = rf.predict_proba(data_input)
|
377 |
+
loan_status = np.array(['N','Y'])
|
378 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
379 |
+
df = pd.concat([data_input, prediction], axis=1)
|
380 |
+
st.write(df)
|
381 |
+
st.subheader(":green[probability] :question:")
|
382 |
+
st.write(pred_proba)
|
383 |
+
|
384 |
+
# Discriminant Analysis
|
385 |
+
if st.sidebar.checkbox("Discriminant Analysis"):
|
386 |
+
st.subheader(":orange[Discriminant Analysis]")
|
387 |
+
lda = pickle.load(open("scoring_lda.pkl", "rb"))
|
388 |
+
pred = lda.predict(data_input)
|
389 |
+
if pred == 1:
|
390 |
+
st.write(":orange[__Le pret peut etre octroyer__] :white_check_mark:")
|
391 |
+
else:
|
392 |
+
st.write(":red[__Desole,...__] :disappointed:")
|
393 |
+
pred_proba = lda.predict_proba(data_input)
|
394 |
+
loan_status = np.array(['N','Y'])
|
395 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
396 |
+
df = pd.concat([data_input, prediction], axis=1)
|
397 |
+
st.write(df)
|
398 |
+
st.subheader(":green[probability] :question:")
|
399 |
+
st.write(pred_proba)
|
400 |
+
|
401 |
+
# XGboost
|
402 |
+
if st.sidebar.checkbox("XGBoost"):
|
403 |
+
st.subheader(":orange[XGBoost]")
|
404 |
+
xg = pickle.load(open("scoring_xg.pkl", "rb"))
|
405 |
+
pred = xg.predict(data_input)
|
406 |
+
if pred == 1:
|
407 |
+
st.write(":orange[__Le pret peut etre octroyer__] :white_check_mark:")
|
408 |
+
else:
|
409 |
+
st.write(":red[__Desole,...__] :disappointed:")
|
410 |
+
pred_proba = xg.predict_proba(data_input)
|
411 |
+
loan_status = np.array(['N','Y'])
|
412 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
413 |
+
df = pd.concat([data_input, prediction], axis=1)
|
414 |
+
st.write(df)
|
415 |
+
st.subheader(":green[probability] :question:")
|
416 |
+
st.write(pred_proba)
|
417 |
+
|
418 |
+
# ANN
|
419 |
+
if st.sidebar.checkbox("Neural Network"):
|
420 |
+
st.subheader(":orange[Neural Network]")
|
421 |
+
ann = pickle.load(open("scoring_ann.pkl", "rb"))
|
422 |
+
pred = ann.predict(data_input)
|
423 |
+
if pred == 1:
|
424 |
+
st.write(":orange[__Le pret peut etre octroyer__] :white_check_mark:")
|
425 |
+
else:
|
426 |
+
st.write(":red[__Desole,...__] :disappointed:")
|
427 |
+
pred_proba = ann.predict_proba(data_input)
|
428 |
+
loan_status = np.array(['N','Y'])
|
429 |
+
prediction = pd.DataFrame(loan_status[pred], columns=['prediction'])
|
430 |
+
df = pd.concat([data_input, prediction], axis=1)
|
431 |
+
st.write(df)
|
432 |
+
st.subheader(":green[probability] :question:")
|
433 |
+
st.write(pred_proba)
|
434 |
+
|
435 |
+
|
436 |
+
|
437 |
+
|
438 |
+
|
439 |
+
# lancer l'application
|
440 |
+
if __name__ == "__main__":
|
441 |
+
main()
|