Predicting medical insurance cost with the help of Machine Learning
In this project the task is to predict charges or cost of the person on the basis of his/her lifestyle, smoking habit, number of childrens and person's home location. From insurance data we can predict future charges of persons, and for predicting the charges of persons we need a strong Machine Learning model to predict charges with good accuracy. This ML model will help insurance company for gaining number of customers because of accurate charges people can protect themselves from insurance scams.
- Created a function which takes data as a input and return basic characteristics of the data
- Visualized the data as we knew the distribution of the data is helpful for better understanding.
- Visualzed the correlation between data, charges distribution of depend on smokers/non-smokers, sex, age, and BMI
- This is a regression problem so I took 3 ML algos: Linear Regression , Random Forest Regressor and XBoost Regressor.
- Choosing right parameters will increase model's accuracy, so tunning models by hyperparameter tunning techniques those are: GridSearchCV and RandomizedSearchCV
- Trained all 3 models and choosed the best one with good score.
- Evaluated that selected model and Visualized the Accuracy.
- Further details are available in notebook.