Students of the Faculty of Computer Science Delivered Reports at PAIS Lab Seminar

The following speakers presented their reports at PAIS Seminar: Rinat Yusupov (Mining causal relations on events from logs), Yana Manukhina (Process mining approach to user interface design), Maxim Savitsky (Performance analysis using software process mining), Alexander Yugov (Discovering service oriented architecture antipatterns using process mining).

Mining causal relations on events from logs

Annotation:
Unceasing efforts to improve processes claim for understanding what is going on inside the process. In this sense, the process mining techniques aim at discovering process behavior from historical recorded data, or event logs. The log structure allows to access to the relations of causal precedence, concurrency, conflict and recently introduced by Stefan Haar et al. revealing between events. Knowledge of these relations facilitates the analysis of a system including partially observable ones, the process insights in the context of diagnosis, or verification; it can also be used to generate more concise representations of behaviors via abstractions.

This work is devoted to the issues of mining causal relations on event from logs as well using these relations in conformance checking and related tasks. We developed an approach to extract the relations and propose a novel method to calculate the fitness of a log using relations.

Keywords: Conformance Checking, Event Log, Event Log Fitness, Event Relations, Occurrence Nets, Petri Nets, Process Mining

Process mining approach to user interface design

Annotation:
Among different factors determining software product's success, usability and user experience are accepted by most as critical. In this presentation I am going to propose a process mining approach to user interface design and analysis. This approach involves applying process mining techniques to event logs of user interaction with a software system and examining the derived process model in order to provide insights into real users' behaviour and motivate usability improvements and software redesign.

Performance analysis using software process mining

Annotation:
In the modern world great attention is paid to information technology. Online Resources introduced in almost every aspect of our lives. Every day, these resources are processed requests from the millions or even billions of people, thus producing a huge amount of records. For the storage and processing of records provided quite productive machines, but this capacity is often not enough. The problem is that the system as a whole contains bottlenecks that require further analysis and improvement.

This work is devoted to the study of varieties bottlenecks in the systems and methods of measurement, constructing a model of the working system based on the obtained records, and identifying bottlenecks for each type of bottleneck.

For data studding was developed a prototype application that allows modeling the system on the basis of submitted entries. The system is designed to process the data, not only static but also dynamic. As a result of the application shows component diagram and the hierarchical model of the whole system, with the possibility of decomposition of individual elements and demonstration of system bottlenecks..

Discovering service oriented architecture antipatterns using process mining

Annotation:
Service oriented architecture is a broadly used paradigm for designing and implementing business collaborations among software applications both within and across organizational boundaries. Once deployed, the system, developed according to SOA, is evolving in two directions: adding new services and redesign of already developed services. As time passes, the quality of the system can be diminishing. In other words, it is necessary to monitor general situation in the system. This paper discusses one of quality characteristics of SOA – incorporation of antipatterns, and provides methods for detection of antipatterns.