Decision trees are popular as stand-alone classifiers or as base learners in ensemble classifiers. Mostly, this is due to decision trees having the advantage of being easy to explain. To improve the classification performance of decision trees, some authors have used Multivariate Decision Trees (MDTs), which allow combinations of features when splitting a node. While there is growing interest in the area, recent research in MDTs all have in common that they do not provide adequate comparison of related work: they do not consider relevant rival techniques, or they test algorithm performance in an insufficient number of databases. As a result, claims have no statistical sustain and, hence, there is a lack of general understanding of the actual capabilities of existing MDT induction algorithms, crucial to improving the state-of-the-art. In this paper, we report on an exhaustive review of MDTs. In particular, we give an overview of 37 MDT induction algorithms, out of which we have experimentally compared 19 of them in 57 databases. We provide a statistical comparison in all databases and subsets of databases according to the number of classes, number of features, number of instances, and degree of class imbalance. This allows us to identify groups of top-performing algorithms for different types of databases.

There is a growing interest in the development of classifiers based on contrast patterns (CPs); partly due to the advantage of them being able to explain classification results in a language that is easy to understand for an expert. CP-based classifiers, when using contrast patterns extracted by miners based on decision trees, attain accuracies comparable with other state-of-the-art classifiers. The existing decision tree-based miners use univariate decision trees (UDTs) to extract CPs. In this paper, we define the concept of multivariate CP. We introduce a multivariate CP miner based on multivariate decision trees (MDTs) as well as a new filtering algorithm for multivariate CPs. From our experimental results, we conclude that our proposed CP miner allows obtaining significantly better classification results than the other state-of-the-art classifiers.

When facing a classification problem, data science practitioners must search through an armoury of methods. Often, practitioners are tempted to use off-the-shelf classifiers, including automated Machine Learning toolboxes; however, stand-alone classifiers are not applicable to any problem and automated Machine Learning may be time-consuming raising up environment-ethical issues. To magnify the problem, (commercial) automated Machine Learning toolboxes are black and practitioners are not allowed to extend them with new methods to improve their classification performance. In this paper, we present a case study, student dropout prediction, which most off-the-shelf classifiers find difficult to solve due to the problem’s inherent class imbalance. We shall see that MS Azure’s autoML outperforms a number of popular, stand-alone classifiers; yet, multivariate PBC4cip, an off-the-shelf classifier especially designed to deal with class imbalance, yields results that are just as good as MS Azure’s autoML. Our studies show that data science practitioners need to build themselves a taxonomy of classification mechanisms in terms of the properties of the problem to solve; additionally, autoML platforms should let scientists modify the armoury of classifiers and provide an explanation of both mechanism selection and mechanism tunning so that practitioners learn further lessons.

Using automated data analysis to understand what makes a play successful in football can enable teams to make data-driven decisions that may enhance their performance throughout the season. Analyzing different types of plays (e.g., corner, penalty, free kicks) requires different considerations. This work focuses on the analysis of corner kick plays. However, the central ideas apply to analyzing all types of plays. While prior analyses (univariate, bivariate, multivariate) have explored the link between contextual factors (e.g., match period, type of defensive marking) and the level of success of a corner kick (e.g., shot, shot on goal, goal), there has been no attempt to combine spatiotemporal event data (sequences of ball movements through the field) and contextual information to determine when and how (strategy) a particular type of corner kick play (tactic) is more likely to succeed or not. To address this gap, we propose an approach that (1) transforms spatiotemporal data into an alternative representation suitable for mining sequential patterns, (2) identifies and characterizes the sequential patterns used by offensive teams to move the ball toward the scoring zone (tactics), and (3) extracts contrast patterns to identify under what conditions different tactics result in increased chances of success or failure; we call these conditions strategies. Our results suggest that favorable and unfavorable conditions for tactic application are not the same across different tactics, supporting the argument that there is a benefit in performing an analysis that treats different tactics separately, where spatiotemporal information plays a crucial role. Unlike prior works on the corner kick, our approach can capture how the interaction between multiple contextual factors impacts the outcome of a corner kick. At the same time, the results can be explained to others in natural languages.