This Data analysis and data visualization project consists of the fact that the average medicine charges can be changed for the smokers and non-smokers.
- Linear Regression
- Polynomial Regression of 2nd degree
- Lasso Regression
- Ridge Regression
- SVM or, Support Vector Machine
- Random Forest Regression
- Decision tree regression
Finally, based on these several algorithms and machine learning methods I have done a comparative analysis of all those algorithms and evaluated them based on the Cross validation Score.
- Distribution of charges for the smokers and non-smokers
- Charges for Men and Women
- BMI Distribution
- Age Distribution of the smokers and non-smokers
- 18+years smokers and nonsmokers
- Joint plot and Scatter Plot showing the distribution
- Distribution of charges of the patients
- Health effects on children due to the parents, who are smoking
- Region based smoker analysis
These features are being evaluated and determined based on the datasets, which provides the exploratory data analysis of this project.
- Error management of various algorithms
- Checking the scores of the algorithms, such as, R2 score, cross validation score, RMSE, mean squared error.
- Evaluation of the best model based on the CV score.
- Evaluation of the best model based on the R2 score.
- Evaluation of the best model based on the RMSE
1.The impact of smoking on medical care use was examined in a 30-month prospective population-based cohort study in Japan (N = 43 408).
2.Male smokers incurred 11% more medical costs than ‘never smokers’ but for female smokers and never smokers the costs were almost thesame.
3.This difference was mainly attributable to the increased use of inpatient medical care among smokers, especially in males, where per monthcost of inpatient care was 33% higher in smokers.