- A Briefly Explain Why A Test And Train Set Were Created As Part Of This Analysis B Explain Briefly What Is Meant By 1 (99.43 KiB) Viewed 33 times
Sample Question. The average career length of a National Basketball Association (NBA) player is close to 5 years1. This means that, on average, players sign two contracts with one or more) team(s) during their career. This happens since their first contract can last up to 3 years and their second contract up to 4 years. During their first year into the NBA, these players are known as rookies. For the majority of the players; their rookie year is considered to be a year of ups and downs where they mainly try to prove themselves to their opponents (and their teammates). For most of the rookies, this is their first time being a part of a professional team and the differences between college basketball (or European basketball) and NBA basketball are huge. The data were split at random into a train and test set (using a ratio of 0.7:0.3) and a classification tree (using conditional inference) was fitted to explore whether the tracking data variables are potentially useful classifiers of whether a basketball player will play more than 5 seasons in the NBA by just looking at their statistics from their rookie seasons. A plot of the classification tree (built using the training data) and the confusion matrix (i.e. a summary of prediction results based on the test set) are provided as output. The variables listed in the tree are games played (GP), points per game (PTS) and number of defensive rebounds (DRB). Confusion matrix. Predicted observed Yes No Sum Yes NO Sum 55 37 92 53 176 229 106 213 319 1 GP p <0.001 566 > 66 2 DREB <0.001 5 PTS p<0.001 59.1 > 9.1 $1.2 >12 6 GP p = 0.037 574 Node 3 (n = 228) Node 4 (n = 189) 1 1 2 2 0.8 0.8 0.6 0.6 0.4 +0.4 0.28 02 0 0 TIT Node 7 (n = 78) Node 8 (n = 124) 1 1 2 2 0.8 0.8 0.6 0.6 0.4 0.4 028 02 0 0 Node 9 (n = 129) 1 2 0.8 0.6 0.4 02 0 Yes