Research Topic: Suppose you're a data analyst at an insurance company. Your role involves estimating the premium each customer should be charged, based on their smoking habits, personal information (age, gender, BMI), family circumstances (number of children), and geographic location. Drawing from the knowledge you've accumulated from Week 1 to Week 6, please analyze the data provided and write a report that is both descriptive and predictive, reflecting your findings. Begin by conducting a succinct descriptive analysis of the data, akin to what you did in your previous assignment. Afterwards, employ your creativity and knowledge to identify the most accurate regression model to forecast insurance charges. Within your regression model, ensure that you shuffle the data initially, then partition it into training data (80%) and testing data (20%). Train your regression model using the training data and subsequently test it (evaluate it) with the testing data. It's crucial to evaluate your regression model using the evaluation metrics we've discussed in class. Guidance for Improving Regression Analysis Performance: For reference, I've conducted a regression analysis on the entire dataset (not divided into training and testing), achieving an R2 value of 0.87 and an RMSE of 4352. In my analysis, I utilized dummy variables to account for categorical variables. Moreover, I used piecewise regression. I separated the data into eight distinct categories and formulated a linear regression for each. These individual linear regressions were then combined into one comprehensive linear regression model using the IF function. My work on this file was limited to one hour. Notably, my analysis solely employed multiple linear regression. Non-linear regression models were not considered. For instance, quadratic regression was not used. If you create better categorizations for your regression models, incorporate non-linear regression models, or performing outlier detection, you may potentially enhance the performance even further.