Siddureddy27/Visualization_Using_Different_Algorithms

1import streamlit as st2import pandas as pd3import numpy as np4import seaborn as sns5import matplotlib.pyplot as plt6from mlxtend.plotting import plot_decision_regions7from sklearn.neighbors import KNeighborsClassifier8from sklearn.linear_model import LogisticRegression9from sklearn.naive_bayes import GaussianNB10from sklearn.metrics import accuracy_score, f1_score11from sklearn.datasets import make_classification,make_circles,make_blobs,make_moons12from sklearn.model_selection import train_test_split,learning_curve13 14st.image("Inno logo.png",width=500)15 16st.title("DB & LC Using Different Algorithms")17st.sidebar.header("Model Selector")18 19data_type = st.sidebar.selectbox("Select Dataset Type", ["Moons","Blobs","Classification","Circles"], index=0)20noise = st.sidebar.slider("Noise Level", 0.0, 1.0, 0.2)21 22# Always Show Number of Neighbors23n_neighbors = st.sidebar.slider("Number of Neighbors", 1, 15, 5)24 25# Algorithm Selection26st.sidebar.subheader("Select Algorithm")27algorithm_selected = st.sidebar.selectbox("Choose an Algorithm", ["KNN", "Logistic Regression","Naive Bayes"])28 29# Show KNN-related parameters only if KNN is selected30if algorithm_selected == "KNN":31    knn_type = st.sidebar.selectbox("KNN Weight Type", ["uniform", "distance"])32    knn_metric = st.sidebar.selectbox("Select KNN Metric", ["euclidean", "manhattan", "minkowski"])33 34# Generate dataset35if data_type == "Blobs":36    X, y = make_blobs(n_samples=2000, centers=2, cluster_std=noise, random_state=27)37elif data_type == "Moons":38    X, y = make_moons(n_samples=2000, noise=noise, random_state=27)39elif data_type == "Circles":40    X, y = make_circles(n_samples=2000, noise=noise, factor=0.5, random_state=27)41else:42    X, y = make_classification(n_samples=2000, n_features=2, n_informative=2, n_redundant=0, n_clusters_per_class=1, random_state=27)43 44# Train-test split and scaling45X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=27)46 47# Train and evaluate selected model48if algorithm_selected == "KNN":49    model = KNeighborsClassifier(n_neighbors=n_neighbors, weights=knn_type, metric=knn_metric)50elif algorithm_selected == "Logistic Regression":51    model = LogisticRegression()52elif algorithm_selected == "Naive Bayes":53    model = GaussianNB()54 55# Train Model56model.fit(X_train, y_train)57y_pred = model.predict(X_test)58 59# Evaluation Metrics60st.write(f"### {algorithm_selected} Evaluation Metrics")61accuracy = accuracy_score(y_test, y_pred)62f1 = f1_score(y_test, y_pred, average='weighted')63 64st.write(f"**Accuracy:** {accuracy:.2f}")65st.write(f"**F1 Score:** {f1:.2f}")66 67# Decision Boundary68st.write(f"### {algorithm_selected} Decision Boundary")69fig, ax = plt.subplots()70plot_decision_regions(X_test, y_test, clf=model, legend=2)71plt.legend(loc='upper right')72st.pyplot(fig)73 74# Learning Curve75st.write(f"### {algorithm_selected} Learning Curve")76train_sizes, train_scores, test_scores = learning_curve(model, X_train, y_train, cv=5, train_sizes=np.linspace(0.1, 1.0, 10))77train_mean = np.mean(train_scores, axis=1)78test_mean = np.mean(test_scores, axis=1)79 80fig, ax = plt.subplots()81ax.plot(train_sizes, train_mean, label='Train Accuracy', marker='o')82ax.plot(train_sizes, test_mean, label='Test Accuracy', marker='s')83ax.set_xlabel("Training Size")84ax.set_ylabel("Accuracy")85ax.legend()86st.pyplot(fig)