Applying the Kaggle dataset, this paper investigates the employment of regression analysis to predict the healthcare expenses based and join them with patient outcomes. A healthcare setting is where this project tests the effectiveness of the obese control methods by means of logistic and linear regressions. While the logistic regression model looked to identify the desired patient outcomes, the linear regression model aimed to determine size of the healthcare costs. Linear model's large mean squared error was received from the tests; it is an aspect that can state that the linear techniques are not competent enough to achieve the required levels of complexity. In contrast, the logistic model scored 100% accuracy strikingly, producing an instance of overfitting. The response to which is presented in this article, emphasizing the requirement for smart modeling strategies and meticulous counterfactual to face better prediction of data in healthcare management. The proposed research first, brings forth salient issues and potential way outs calls for still deeper study and research as well as demonstrating the efficacy of regression analysis in the medical field.
• Age Distribution: This histogram demonstrates the number of the patients in the group of every age range below. The figure of the patient population on the general age profile is enhanced by two methods superimposed on the histogram. One of them is density estimation presented as smoother line, the other is real histogram frequency.
• Heatmap is an example of visualized correlation between such numeric data as age, room number bed, billing amount and time of stay. It becomes easier to look out for any relationship or association between the variables that may affect feature engineering and model accuracies with this illustration's help.
• The plot shows variation in the residuals, with a particular pattern of the spread growing from its minimum value towards a maximum for the value of y. Thus, to the contrary of linear regression, linear regression modes need to be modified disposing linear input feature transformations as one of the essential linear regression principles.