Because of the FIRST® Robotics Competition (FRC) World Championship, I'm in Houston. Since I arrived several days before the rest of my team, I met and am going to meet several Rice professors about data science.
After talking with Prof. Jermaine about my project, I learned quite a bit about what I could do to expand the scope of my project. Dr. Chris Jermaine is the head of the Rice Computer Science Department and the first professor I met. My project covers the Thucydides Trap in which a rising power and a ruling power come into conflict, in 12/16 cases there is a war. Using this information the outcome of the U.S.A/China conflict can be predicted. One of the main issues with the current model is that it only accounts for time, ignoring other historical elements which affected the outcome. For example, the main reason the Cold War never broke out into a direct war between the U.S. and U.S.S.R is the threat of nuclear Mutually Assured Destruction, something that did not exist in previous conflicts.
Since Prof. Jermaine had worked with situations in which large data sets were analyzed through hundreds of parameters, he is familiar with many different models which could be used to illustrate the effects of different factors. He suggested I use a Logistic Regression model to illustrate non-time parameters. This is particularly effective because the input of the function is continuous but the output is binary(ex. War vs. No war). Another important thing Prof. Jermaine mentioned was to make sure that I do not do too many parameters since the data would have to be hand collected, increasing the scope exponentially.
It was both nerve-racking and interesting to talk to my first professor in data science.