Add application files

.streamlit/config.toml

@@ -0,0 +1,5 @@
+[theme]
+base="dark"
+primaryColor="#FF3333"
+backgroundColor="#3d0404"
+secondaryBackgroundColor="#1e0203"

README.md

@@ -1,12 +1,4 @@
----
-title: Stroke Prediction
-emoji: 💩
-colorFrom: gray
-colorTo: gray
-sdk: streamlit
-sdk_version: 1.10.0
-app_file: app.py
-pinned: false
----
+# Stroke Prediction using Machine Learning
+The aim of this project is to develop a model which predicts whether a patient is likely to get a stroke based on the parameters like gender, age various diseases and smoking status.
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Dataset used: https://www.kaggle.com/datasets/fedesoriano/stroke-prediction-dataset

app.py

@@ -0,0 +1,26 @@
+import streamlit as st
+from multiapp import MultiApp
+from apps import home, pred, data, model
+
+st.set_page_config(page_title='Stroke Prediction using ML - Mini-Project for 19CS601', page_icon = 'favicon.png', initial_sidebar_state = 'auto')
+
+# Hide Streamlit brandings
+hide_streamlit_style = """
+            <style>
+            #MainMenu {visibility: hidden;}
+            footer {visibility: hidden;}
+            </style>
+            """
+st.markdown(hide_streamlit_style, unsafe_allow_html=True)
+
+app = MultiApp()
+
+app.add_app("Home", home.app)
+app.add_app("Prediction Service", pred.app)
+app.add_app("Dataset Overview", data.app)
+app.add_app("Model Overview", model.app)
+
+with st.sidebar:    
+    sess = app.run()
+
+app.view(sess)

apps/data.py

@@ -0,0 +1,23 @@
+import streamlit as st
+import pandas as pd
+
+def app():
+    with st.sidebar:
+        st.title('Stroke Prediction using Machine Learning')
+
+        st.write('This model which predicts whether a patient is likely to get a stroke based on the parameters like gender, age various diseases and smoking status.')
+        st.markdown('_For Machine Learning - 19CS601_')
+    
+    st.title('Dataset Overview')
+
+    st.write("The following is the DataFrame of the healthcare dataset for stroke prediction.")
+    st.write('This dataset is used to predict whether a patient is likely to get stroke based on the input parameters like gender, age, various diseases, and smoking status. Each row in the data provides relavant information about the patient.')
+
+    st.markdown('Dataset by Federico Soriano Palacios ([__fedesoriano__](https://www.kaggle.com/fedesoriano) on Kaggle)')
+    st.markdown('Source: https://www.kaggle.com/datasets/fedesoriano/stroke-prediction-dataset')
+
+    df = pd.read_csv("dataset\healthcare-dataset-stroke-data.csv")
+    df['hypertension'] = df['hypertension'].map({0:"No", 1:"Yes"})
+    df['heart_disease'] = df['heart_disease'].map({0:"No", 1:"Yes"})
+    df['stroke'] = df['stroke'].map({0:"No", 1:"Yes"})
+    st.write(df)

apps/home.py

@@ -0,0 +1,15 @@
+import streamlit as st
+from PIL import Image
+
+def app():
+    with st.container():
+        st.title('Stroke Prediction using Machine Learning')
+        st.markdown('For _Machine Learning - 19CS601_')
+
+        st.write('This model which predicts whether a patient is likely to get a stroke based on the parameters like gender, age various diseases and smoking status.')
+
+        st.write('* Pick the \'Prediction Service\' to check the working of the model.')
+
+        st.write('* Pick the \'Dataset Overview\' to know more about the dataset.')
+
+        st.write('* Pick the \'Model Overview\' to know more about the model that we have used for predictions.')

apps/model.py

@@ -0,0 +1,54 @@
+import streamlit as st
+import pickle
+import lightgbm
+from sklearn.metrics import classification_report,plot_precision_recall_curve,plot_confusion_matrix,precision_recall_fscore_support,plot_roc_curve
+
+def app():
+    with st.sidebar:
+        st.title('Stroke Prediction using Machine Learning')
+
+        st.write('This model which predicts whether a patient is likely to get a stroke based on the parameters like gender, age various diseases and smoking status.')
+        st.markdown('_For Machine Learning - 19CS601_')
+    
+    st.title('Model Overview')
+    st.write('The model performance of the dataset is presented below.')
+
+    # Retreving model and it's components for performance metric
+    model = pickle.load(open("apps\models\gbm\gbm-model-pickle.sav", 'rb'))
+    X_test = pickle.load(open("apps\models\gbm\gbm-xtest.sav", 'rb'))
+    Y_test = pickle.load(open("apps\models\gbm\gbm-ytest.sav", 'rb'))
+    Y_pred = model.predict(X_test)
+
+    st.header('Model performance')
+    #result = model.score(X_test, Y_test)
+
+    precision,recall,f1_sc,support=precision_recall_fscore_support(Y_test,Y_pred)
+    accuracy=model.score(X_test,Y_test)
+
+    col1, col2, col3, col4 = st.columns(4)
+    col1.metric("Accuracy", round(accuracy,4), "")
+    col2.metric("Recall", round(recall[0],4), "")
+    col3.metric("F-measure", round(f1_sc[0],4), "")
+    col4.metric("Support", support[0], "")
+
+    st.subheader("Model type: ")
+    st.write(model)
+
+    st.set_option('deprecation.showPyplotGlobalUse', False)
+    st.subheader("Confusion Matrix: ")
+    plot_confusion_matrix(model, X_test, Y_test, display_labels=['NoStroke','Stroke'])
+    st.pyplot()
+    #st.table(confusion_matrix(Y_test, Y_pred))
+
+    st.subheader("ROC Curve")
+    plot_roc_curve(model, X_test, Y_test)
+    st.set_option('deprecation.showPyplotGlobalUse', False)
+    st.pyplot()
+
+    st.subheader("Precision-Recall Curve")
+    plot_precision_recall_curve(model, X_test, Y_test)
+    st.pyplot()
+
+    st.subheader('Other metrics:')
+    report=classification_report(Y_test, Y_pred, target_names=None)
+    st.code(report)

apps/pred.py

@@ -0,0 +1,113 @@
+import streamlit as st
+import lightgbm
+import pickle
+import numpy as np
+from sklearn.preprocessing import RobustScaler
+from sklearn.decomposition import PCA
+model = pickle.load(open("apps\models\gbm\gbm-model-pickle.sav", 'rb'))
+scaler = pickle.load(open("apps\models\gbm\gbm-scaler.sav", 'rb'))
+
+def app():
+    with st.sidebar:
+        st.title('Stroke Prediction using Machine Learning')
+
+        st.write('This model which predicts whether a patient is likely to get a stroke based on the parameters like gender, age various diseases and smoking status.')
+        st.markdown('_For Machine Learning - 19CS601_')
+
+        st.write('It may take a few moments to complete this survey.')
+
+    with st.container():
+        st.subheader('Stage 1: Personal Questions')
+
+        ch_gender = st.selectbox(
+            'Gender: ',
+            ('Male', 'Female', 'Others'))
+            
+        ch_age = st.number_input('Age: ',min_value=0, max_value=150, value=18,step=1)
+
+        ch_restype = st.radio(
+            'Residence Type: ',
+            ('Urban', 'Rural'))
+
+        ch_marital = st.radio(
+            'Did you ever get married? ',
+            ('Yes', 'No'))
+
+        ch_worktype = st.selectbox(
+        'Work type: ',
+        ('I\'m a child.', 'I\'m self employed', 'Working for the Private.','Working for the Government.','Never worked for anyone.'))
+
+        st.subheader('Stage 2: Health Questions')
+
+        ch_height = st.number_input('Height (in m): ',min_value=0.0, max_value=500.0, value=175.0,step=0.1)
+
+        ch_weight = st.number_input('Weight (in kg): ',min_value=0.0, max_value=5000.0, value=75.0,step=0.01)
+        
+        calc_bmi  = ch_weight / (ch_height/100)**2
+
+        ch_bmi = st.number_input('BMI: (Optional)',min_value=0.0, max_value=60.0, value=calc_bmi,step=0.01)
+
+        ch_agl = st.number_input('Average Glucose Level (in mg/dL): ',min_value=50.0, max_value=300.0, value=50.0,step=0.01)
+
+        ch_smokingstat = st.selectbox(
+            'Smoking status: ',
+            ('Never smoked', 'Formerly smoked', 'I\'m an active smoker','I prefer not to speak'))
+
+        st.write('Are you currently suffering from these diseases?')
+
+        ch_hypertn = st.checkbox('Hypertension')
+
+        ch_hearttn = st.checkbox('Heart Disease')
+
+        submit = st.button('Submit')
+
+        if submit:
+            
+            ch_gender = 0 if ch_gender=="Female" else 1 if ch_gender=="Male" else 2
+            ch_marital =  1 if ch_marital=="Yes" else 0
+            ch_worktype = 1 if ch_worktype=="Never worked for anyone." else 4 if ch_worktype=="I\'m a child." else 3 if ch_worktype=="I\'m self employed" else 2 if ch_worktype=="Working for the Private." else 0
+            ch_restype = 1 if ch_restype=="Urban" else 1
+            ch_smokingstat = 3 if ch_smokingstat=="I\'m an active smoker" else 1 if ch_smokingstat=="Formerly smoked" else 2 if ch_smokingstat=="Never smoked" else 0
+            ch_hypertn =  0 if ch_hypertn==False else 1 if ch_hypertn==True else 999
+            ch_hearttn =  0 if ch_hearttn==False else 1 if ch_hearttn==True else 999
+
+            input = scaler.transform([[ch_gender,ch_age,ch_hypertn,ch_hearttn,ch_marital,ch_worktype,ch_restype,ch_agl,ch_bmi,ch_smokingstat]])
+
+            prediction = model.predict(input)
+            predictval = model.predict_proba(input)
+            
+            with st.expander("Results"):
+                if prediction==0:
+                    str_result = 'The model predicts that with the probability of %.2f%%, you won\'t be suffering from stroke in the future.'%(predictval[0][0]*100)
+                    st.success(str_result)
+                    st.write("""
+                        The best way to help prevent a stroke is to eat a healthy diet, exercise regularly, and avoid smoking and drinking too much alcohol.
+                        These lifestyle changes can reduce your risk of problems like:
+                        - arteries becoming clogged with fatty substances (atherosclerosis)
+                        - high blood pressure
+                        - high cholesterol levels
+                        If you have already had a stroke, making these changes can help reduce your risk of having another stroke in the future.
+
+                    """)
+                    st.write("Source: [National Health Service (NHS) - United Kingdom](https://www.nhs.uk/conditions/stroke/prevention/)")
+
+                elif prediction==1:
+                    str_result = 'The model predicts that with the probability of %.2f%%, you will be suffering from stroke in the future.'%(predictval[0][1]*100)
+                    st.error(str_result)
+                    if predictval[0][1] >= 0.85:
+                        st.subheader("Please seek medical attention as early as possible to mitigate the stroke disease.")
+                    st.write("""
+                        The best way to help prevent a stroke is to eat a healthy diet, exercise regularly, and avoid smoking and drinking too much alcohol.
+                        These lifestyle changes can reduce your risk of problems like:
+                        - arteries becoming clogged with fatty substances (atherosclerosis)
+                        - high blood pressure
+                        - high cholesterol levels
+                        If you have already had a stroke, making these changes can help reduce your risk of having another stroke in the future.
+
+                    """)
+                    st.write("Source: [National Health Service (NHS) - United Kingdom](https://www.nhs.uk/conditions/stroke/prevention/)")
+
+                else:
+                    st.error('NaN: Unexpected error')
+                    st.markdown("Debug: Selected input:")
+                    st.code(input)

multiapp.py

@@ -0,0 +1,23 @@
+"""Frameworks for running multiple Streamlit applications as a single app.
+"""
+import streamlit as st
+
+class MultiApp:
+    def __init__(self):
+        self.apps = []
+
+    def add_app(self, title, func):
+        self.apps.append({
+            "title": title,
+            "function": func
+        })
+
+    def run(self):
+        app = st.selectbox(
+            'Navigation',
+            self.apps,
+            format_func=lambda app: app['title'])
+        return app
+
+    def view(self, app):
+        app['function']()

requirements.txt

@@ -0,0 +1,5 @@
+streamlit==0.71.0
+pandas==1.1.3
+scikit-learn==0.23.2
+numpy==1.19.2
+lightgbm==3.3.0