Sorry, all you Game of Thrones fans — Milan Janosov, a Ph.D. candidate at the Central European University Center for Network Science, thinks the Mother of Dragons is “quite likely to die,” with other popular characters like Arya Stark and the Hound likewise in “dangerous positions.”
Why? Science! Data science, to be more precise.
Janosov has built a quantitative prediction model showing which Game of Thrones characters are statistically likely to die next, something that he says started as a class project for a data mining lecture. He created what amounts to a social network for the show, as he explains in a report about his model: “Game of Thrones is a complex world in which social position and true friends seem to be quite important, so I quantified each character’s social interaction patterns using the tools of network science. I then predict their fate using machine learning methods.”
Here’s his hypothesis, as he explains it to BGR:
After building what you could call the social network — plotting out each individual character, and studying the linkages between them — Janosov then used what he says is a well-known machine learning approach (the support vector machine model) to calculate the probability of which still-living characters will die.
“The scientific part is the question: seemingly the friends, enemies and allies are super important in the Game of Thrones realm,” he tells BGR. “Does this mean that the various measures of this — centrality measures in the characters’ social network — are good predictors of a character’s death?
“The gist of the calculation is that we consider all the dead characters, and see what their social collaboration patterns look like — and based on how much alike a still living character is, the higher her chances are to die. So this is a hypothesis, and during the upcoming episodes we’re going to learn how well … the various measures of social position (are good predictors).”
Janosov is starting his second year in the network science PhD program at the Central European University in Budapest. The parts about the Game of Thrones universe he especially loves, Janosov explains, are the complexity, the level of detail and the unexpected plot twists.
All of it also making for abundant fodder for a data science student keen on graphs and connections and patterns.
There are 94 characters interesting enough for Game of Thrones fans to care about, Janosov explains in his report. Sixty-one characters have already died at the time of this writing.
Based on that knowledge, he writes, you could form “an educated guess” of who’s going to kick the bucket next by looking at “which of those people still alive have similar features to those who have already passed away? … This problem resembles the well-known churn problem, which can be solved with various classification-based algorithms.”
Looking at the network centrality patterns — and the error of those probabilities, based on repeating the prediction “with five-fold cross-validation” a hundred times — Janosov’s model comes up with a 91 percent chance that Daenerys Targaryen will meet an untimely end soon.
Here’s a table from his report stacking who’s likely to die next, from the least likely at the bottom all the way up to most likely:
“The list tells us many interesting things,” Janosov writes. “First, Daenerys seems to be quite likely to die, overlapping with a number of speculations, while Tyrion and Jon Snow seem to be relatively safe. Second, both the ever-popular Arya Stark and the less friendly Hound, already so close to death many times before, are both in dangerous positions. Surprisingly, Cersei, currently sitting on the Iron Throne, and Baelish who is doing his best to get there, seem to be in a much better position. It seems Jorah Mormont will find the cure for his greyscale disease, and despite all he has been through, Theon Greyjoy will probably survive. Sadly, the same cannot be said about the Arryn family.”
This project, meanwhile, is similar to one performed by a team from the Technical University of Munich that was a product of a JavaScript course there. That team similarly used comparisons between characters who have died and still-living characters to see if there are tell-tale signs that suggest who will die next.
For what it’s worth, here’s that team’s ranking: https://www.got.show/ranking. It similarly assigns Daenerys Targaryen a high probability of dying soon, and it should be noted — this team has already been correct with the number one and two names on its list of who’s dying next — Tommen and Stannis Baratheon. (Daenerys is third on the list.)
