Project of statistic about modelling cancer causalities. This was made in collaboration by Lorenzo Tarricone and Karim Abdel Sadek
This assignment was given by our statistic professor as compulsory project and we decided to push it foreward to try to come out with the best project we could.
In this ambitious work we tried to answer a difficult question that is: "What is causing cancer?" . Of course we don't have the instruments to give an answer, but still we attempted to adress this question with a fundamental statistical tool that is MULTIPLE LINEAR REGRESSION.
We proceeded by selecting some risk factors and then by trying to find correlation between those and the number of people suffering of cancer. The research was brouth out on more than 120 countries and considering more or less the last 15 years.
Given the first raw model we proceeded with different types of model selction techniques to try to see if our results were roboust and consistent.
So: does smoking really causes cancer (according to our data and keeping in mind that correlation doesn't imply causation)? Read the research to find it out!
P.S. Don't forget to check out the HTML VERSION