Research on an improved prediction model based on decision tree algorithm

In this research, we collect and complete a data set of freshmen registration, and use the machine learning algorithm of decision tree to predict it. Our goal is to use this algorithm to verify whether the enrollment of freshmen can be predicted. Then we adopt the pruning scheme to optimize the decision tree. One is to prune the number of layers of the tree, and the other is to prune the number of samples on the leaf nodes. Then, we present the confusion matrix and F1 score to assess the impact of machine learning. The exploratory comes about to appear that the pruning scheme has greatly improved our prediction model.