Best Insurance Claim Prediction Web App Using Machine Learning

Insurance Claim Prediction

OverviewThe Insurance Claim Prediction App is a web tool made with Streamlit that uses machine learning to guess how much someone’s insurance claim might be. It looks at stuff like a person’s age, health, and other personal details to make its predictions.It’s built using a trained regression model, which basically means it learned from real insurance data. The app gives quick and pretty accurate results, which can help insurance companies plan better and make smarter money decisions.

Project Information

Download New Real Time Projects :-Click here

Business Problem

Insurance companies need reliable tools to predict claim amounts for effective resource allocation, pricing policies, and financial forecasting. This application solves that problem by providing a regression-based predictive model that takes user input and delivers estimated claim amounts.

By incorporating demographic and health-based features such as age, BMI, smoker status, and more, the system provides accurate forecasts tailored to each individual.

Dataset Description

The dataset used for training includes the following key features:

• Age
• Gender
• Body Mass Index (BMI)
• Smoking status
• Diabetic status
• Number of children
• Region
• Blood pressure

The target variable is a numerical value that represents the insurance claim amount.

Features Included

• Streamlit-based User Interface: Clean and interactive web interface for users to input details.
• Trained ML Model: Uses a regression model saved in insurance.pkl to predict claim values.
• Custom Input Form: Allows dynamic user inputs for age, gender, health conditions, etc.
• One-click Prediction: Users simply click a button to generate predictions.
• Interactive App Layout: Designed for usability and clarity using Streamlit’s layout tools.

Project Files

• app.py: Main Python file that powers the Streamlit web application.
• insurance.pkl: Pickled file that contains the pre-trained machine learning regression model.
• requirements.txt: Contains all necessary Python libraries for the application to run.

Model Training & Deployment Process

Data Preprocessing

• Encoded categorical values (e.g., gender, smoker)
• Handled missing values
• Normalized/Scaled numeric features where necessary

Exploratory Data Analysis (EDA)

• Analyzed correlations between features and claim amount
• Identified outliers and adjusted data accordingly

Feature Selection

• Selected most influential features using correlation and model testing

Model Building

• Tested multiple regression algorithms such as:
  • Linear Regression
  • Decision Tree Regressor
  • Random Forest Regressor
  • Gradient Boosting Regressor
• Selected the best-performing model based on evaluation metrics like:
  • Mean Absolute Error (MAE)
  • Mean Squared Error (MSE)
  • Root Mean Squared Error (RMSE)

Deployment

• Final model saved using Pickle
• Integrated with Streamlit for web access and prediction

App Interface

The web application allows users to enter the following fields:

• Age
• Gender
• BMI
• Blood Pressure
• Diabetic status
• Number of children
• Smoking status
• Region

Once all inputs are filled in, the user clicks the Predict button to get the estimated insurance claim amount instantly displayed on the interface

How to Use

1. Extract the files
   Ensure all project files are in the same directory.
2. Install dependencies
   Use the following command: pip install -r requirements.txt
3. Start the web application streamlit run app.py
4. Open in your browser
   Visit: http://localhost:8501

We have projects Available in all languages:–Click Here