Abstract:

Growing IT complexity and related problems which are converted into IT tickets create a need for new qualitative approaches to automate the extraction of main topics described in those tickets with the goal to provide high quality support for the IT process workers while enabling a smooth service delivery to the end customer. Present paper proposes a method of knowledge extraction in the form of stylistic patterns in business process (BP) textual data, here in IT tickets, with the goal to identify their Decision-Making Logic (DML) level and the implied ticket complexity. The results of experimental analysis of a data set of ticket textual data from the ITIL-based Change Management process showed that specificity of stylistic patterns depends on the proposed feature of the ticket length, which in its turn reveals a positive correlation with the DML levels implying the ticket complexity.