Συμμορφωτική πρόβλεψη ακολουθιακών δεδομένων με δένδρα απόφασης
Conformal prediction of sequential data with decision trees
The purpose of this paper is to deal with the study of conformal prediction with the use of some decision trees. More specifically, it aims at applying the method to a problem of prediction the final outcome of football matches for two specific teams, using the corresponding categorization trees. Two sets of sequential data (one for each team) were used in the implementation of the method. The first dataset refers to Arsenal’s series of matches this season (2018-2019) by March 20 and consists of 30 observations, while the second dataset refers to Barcelona’s matches in the same season and consists of 27 observations. The conclusion that emerged after the use of the corresponding categorization algorithm is that the predictive model with the use of categorization trees has difficulty in predicting the final outcome of the matches, indicating the degree of difficulty for safe predictions in the field of football.