Use rpart to build a decision tree based on the last 1000 instances from the given dataset. Train the model on only the descriptive features, but excluding ID. Set the parameters as follows: complexity parameter should be 0.01, the minimum split size should be 2, the minimum leaf size should be 1, and the maximum tree depth should be 5. Use this tree to make a prediction on the instance represented by the following vector (only the target feature is excluded from this vector): (10001,B,Flight,3,4,2412,3,high,F,6,5.68).