md-k-sarker/predicting-health-insurance-cost

Predicting health insurance cost from Morality data using Machine Learning techniques

Can the cost of health insurance be estimated automatically?

Background:

Insurance company especially health insurance company need to analyze how a customer is disease prone. We will try to find the risk based on the customers data.

If the Health_Risk can be estimated accurately then insurance cost can be calculated using this formula.

Health Insurance Cost = Base Cost + α * Health Risk

We try to investigate how accurately we can estimate the health risk of a person.

Data Source: National Center for Health Statistics

Data Preprocessing

Total records : 1.04 M Has disease and non disease records. Disease related records : 0.56 M No. of feature: 34

Feature information:

Median and Quartile of each features

Outlier Detection

Used Box-Whisker plot to detect outlier.

Class Imbalance

Has high class imbalance

Classification

We applied K-Nearest Neighbour, Random Forest, Neural Network and Support Vector Machine algorithms to classify the different health risk.

predicting-health-insurance-cost's Issues

morality typo

Your repo description uses the word "morality" but it should be "mortality". I saw the same typo in a few filenames, but something to watch out for.