Code

import pandas as pd
from sklearn.neighbors import NearestNeighbors
from sklearn.metrics.pairwise import cosine_similarity
# Load the dataset
dataset = pd.read_csv('hinsurance.csv').rename(columns=lambda x: x.strip())
# Replace with the path to your dataset file
# Preprocess the 'Premium' column
dataset['Premium'] = dataset['Premium'].str.replace('Rs. ', '').str.replace(',', '').astype(float)
dataset['Deductible'] = dataset['Deductible'].str.replace('Rs. ', '').str.replace(',', '').astype(float)
dataset['Maximum out-of-pocket'] = dataset['Maximum out-of-pocket'].str.replace('Rs. ', '').str.replace(',', '').astype(float)
dataset['Income'] = dataset['Income'].str.replace('Rs. ', '').str.replace(',', '').astype(float)
# Preprocess the 'Deductible' column
dataset['Coinsurance'] = dataset['Coinsurance'].str.replace('Rs.', '').str.replace(',', '').str.strip()
dataset['Coinsurance'] = dataset['Coinsurance'].apply(lambda x: '0' if not x.isdigit() else x).astype(float)
# Drop rows with missing values
dataset = dataset.dropna()

# Encode categorical features
dataset_encoded = pd.get_dummies(dataset, columns=['Type', 'Coverage', 'Gender', 'Health status', 'Occupation'])

# Manually provide plan_name column
plan_name_column = pd.DataFrame({'plan_name': ['Aarogya Plus', 'Bharti AXA Health Plus', 'Max Bupa Family Health Optima',
    'Star Health Optima Health Plus', 'Apollo Munich Health Optima', 'Religare Health Optima',
    'ICICI Lombard Health Optima', 'HDFC Ergo Health Optima', 'Bajaj Allianz Health Optima',
    'Nippon Life Health Optima', 'Kotak Mahindra Health Optima', 'SBI General Health Optima',
    'New India Assurance Health Optima', 'HDFC Ergo Health Plus', 'Tata AIG Health Insurance',
    'Oriental Insurance Health Plus', 'United India Health Optima', 'Future Generali Health Protect',
    'HDFC ERGO Health Premier', 'Aditya Birla Activ Assure Diamond', 'National Insurance Mediclaim Plus',
    'Reliance HealthGain', 'Oriental Insurance Happy Family Floater', 'HDFC Ergo Health Suraksha Gold',
    'Star Health Senior Citizen Red Carpet', 'United India Family Medicare Gold', 'Max Bupa Health Recharge',
    'SBI General Arogya Premier', 'ICICI Lombard Complete Health', 'Bajaj Allianz Health Guard',
    'Reliance General Wellness', 'New India Assurance Mediclaim', 'HDFC Ergo My:Health Suraksha',
    'Star Health Family Optima', 'Oriental Insurance Health Protect', 'United India Health Protector',
    'Max Bupa Health Companion', 'SBI General Arogya Plus', 'Apollo Munich Optima Restore',
    'Bajaj Allianz Family Floater', 'HDFC Ergo Health Suraksha Diamond', 'Star Health Comprehensive',
    'ICICI Lombard Health Shield', 'Reliance Health Insurance', 'Oriental Insurance Health Protect Plus',
    'United India Super Top-up', 'Max Bupa Health Premia', 'SBI General Retail Health',
    'New India Assurance Mediclaim Plus', 'HDFC Ergo Optima Super', 'Religare Care Freedom',
    'Star Health Senior Citizen Red Carpet', 'ICICI Lombard Health Protect Plus',
    'Bajaj Allianz Health Infinity', 'Max Bupa Health Premia Gold', 'SBI General Arogya Premier',
    'United India Individual Mediclaim', 'Oriental Insurance Health of Family',
    'New India Assurance Family Floater']})

# Concatenate plan_name column with encoded features
dataset_encoded = pd.concat([plan_name_column, dataset_encoded], axis=1)

# Extract the feature vectors and target variable
target_column = 'plan_name'
X = dataset_encoded.drop(columns=target_column).values

# Compute the similarity matrix
similarity_matrix = cosine_similarity(X)

# Fit the KNN model
knn_model = NearestNeighbors(metric='cosine')
knn_model.fit(X)

# Get user input
user_input = {
    'Type': input('Enter the insurance type: '),
    'Coverage': input('Enter the desired coverage: '),
    'Premium': float(input('Enter the maximum premium you are willing to pay: ')),
    'Deductible': float(input('Enter the maximum deductible you are willing to pay: ')),
    'Coinsurance': float(input('Enter the maximum coinsurance percentage you are willing to pay: ')),
    'Maximum out-of-pocket': float(input('Enter the maximum out-of-pocket expense you are willing to pay: ')),
    'Age': int(input('Enter your age: ')),
    'Gender': input('Enter your gender: '),
    'Health status': input('Enter your health status: '),
    'Income': float(input('Enter your income: ')),
    'Occupation': input('Enter your occupation: ')
}
numerical_features = ['Premium', 'Deductible', 'Coinsurance', 'Maximum out-of-pocket', 'Income', 'Age']

# Create a target vector for user input
target_vector = []
for feature in dataset_encoded.columns:
    if feature == target_column:
        continue
    if feature in numerical_features:
        std_dev = dataset_encoded[feature].std()
        if std_dev == 0:
            target_vector.append(0)
        else:
            target_vector.append((user_input.get(feature, 0) - dataset_encoded[feature].mean()) / std_dev)
    else:
        target_vector.append(1 if (feature + '_' + user_input.get(feature, '')) in dataset_encoded.columns else 0)
target_vector = [target_vector]# Compute similarity between user input and existing plans
user_similarity_scores = cosine_similarity(target_vector, X)[0]

# Find the most similar plan to the user input
most_similar_plan_index = user_similarity_scores.argmax()
def recommend_insurance_plan(dataset, similarity_matrix, plan_index, top_n=5):
    similarity_scores = similarity_matrix[plan_index]
    similar_plans_indices = similarity_scores.argsort()[::-1][1:top_n + 1]
    return dataset.iloc[similar_plans_indices]
# Get recommendations based on the most similar plan
recommended_plans = recommend_insurance_plan(dataset, similarity_matrix, most_similar_plan_index, top_n=5)
print(dataset.loc[recommended_plans.index, ['plan_name', 'Coverage', 'Premium', 'Deductible']])