Predicting Winners of Big Brother Using Predictive Methods of Data Mining

Abstract

Data mining techniques such as classification methods can be used to predict the value of a variable based on other attributes in the same dataset. Prediction models can be used on data like sporting statistics and political polls. One area that prediction has not crossed over into yet is reality television. These shows have winners, many seasons, lots of contestants, and even more information about the players released to the public before the show starts. A great example of this is Big Brother. Before every season, each contestant completes pre-season interviews with the producers, and these answers are released to the public. This information, along with their demographic information, could be used to predict the winner before one player steps foot into the house. Ideally, a dataset would already exist with all this information on each contestant of Big Brother, but it does not. Therefore, the first step needs to be creating this dataset. After this happens, testing can begin. This dataset would be tested with prediction models, specifically classification methods. Once the testing begins, any data revision can occur. For example, if there is not enough data, information about contestants from Big Brother Canada and Big Brother UK can be added to the dataset. Hopefully, in the end, a model will be able to correctly identify a winner, or future winner, of Big Brother.