Predictive monitoring of business processes within the framework of digital economy development

Methods. In this paper, we present an approach to analyze such event logs in order to predictively monitor business constraints during business process execution. At any point during an execution of a process, the user can define business constraints in the form of linear temporal logic rules. The framework defines the input data values that are more or less likely to meet each business constraint. The approach has been implemented in the ProM process mining toolset. Results. The results show that this model can provide competitive classification performance, create highly interpretable models, and effectively reduce data preparation efforts. The calculation of the information gain coefficient criteria is performed and shown in the algorithm using the appropriate equations and a recursive function. In this paper, we describe J48SS, a new decision tree inducer based on the C4.5 algorithm. The algorithm is experimentally validated on a real business speech analytics setting. J48SS is shown to effectively reduce the data preparation effort, and the use of binary splits shows that the tree grows fully. Future work could be devoted to investigating ensembles of J48 trees and applying the proposed algorithm to a corresponding database, where all types of supported attributes naturally arise. Also, a future research direction could be to extend the model to deal with temporal logic formulas. Such a formulation would allow the decision tree to take into account relationships between the values of different attributes, instead of looking at each of them individually. Novelty. This paper presents J48SS, a new decision tree learner that can blend static, sequential, and time-series data for classification purposes. The new algorithm is based on the popular C4.5 decision tree learner and relies on the concepts of frequent pattern extraction and timeseries formula generation. A framework for predictive business process monitoring is shown. Practical value. In this article, we propose implementation of the MapReduce C4.5 algorithm. Empirical results indicate that the implementation of the algorithm demonstrates both time efficiency and scalability.