Erratum: Process Mining to Discover Shoppers' Pathways at a Fashion Retail Store Using a WiFi-Base Indoor Positioning System(IEEE Transactions on Automation Science and Engineering(2017) 14: 4(1786–1792)Doi:10.1109/TASE.2017.2692961)

Cong Liu, Qingtian Zeng, Mengchu Zhou

Research output: Contribution to journalComment/debatepeer-review

Abstract

In the October 2017 issue of the IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING [1], there are several errors, reported and corrected as follows. Comment 1: There are errors in [1, Definition 2]. Definition 2 (Petri Nets): [1] A Petri net is a process model that uses modeling formalism to represent the flow of a process. In this paper, we used a Petri net to provide information about a discovered model. A Petri net is a triplet (P, T, F), where P is a set of places and T is a set of transitions such that P × T = ø, and F is a set of directed arcs defined as a Cartesian product of P and T: F = (P × T) U (T × P). It should be revised as follows. Definition 2 (Petri Nets): A Petri net is a process model that uses a modeling formalism to represent the flow of a process. It is a triplet (P, T, F), where P is a set of places and T is a set of transitions such that P × T = ø, and F (P × T) U (T × P) is a set of directed arcs from places to transitions and from transitions to places. Note that arcs of a Petri net are defined as a subset of the Cartesian product of P and T, i.e., F (P × T) U (T × P) [2]. Comment 2: The following statement on the first line of the right column of [1, p. 1788] is wrong. "A workflow net is a derivative of the Petri net that guarantees model soundness." A workflow net is defined as a Petri net with a dedicated source place where the process starts and a dedicated sink place where the process ends. However, a workflow net cannot always be guaranteed to be sound. The soundness of a workflow net is decided by checking the boundedness and liveness properties of its corresponding short-circulated net [3]. Comment 3: [1, Fig. 1(a)] is incorrect because it fails to highlight all block structures of the process tree in Fig. 1(b). The correct one is shown in Fig. 1 of this letter. Comment 4: [1, Fig. 2] is incorrect as it does not match the presented log. The corrected one is shown in Fig. 2 of this letter. Comment 5: The following statement on the 19th line of the right column on p. 1790 is not correct. "We used an inductive miner with ProM 6.5.2 and observed that the models are completely different, as shown in Figs. 7 and 8." The process models in Figs. 7 and 8 are not completely different. On the contrary, these two models describe very similar behavior if we consider their produced languages/sequences. Here we take the incoming place of 1f-right as the initial marking and the place without outgoing arcs as the final marking for both models. More precisely, almost all possible sequences produced by the model in Fig. 7 can be replayed by the model in Fig. 8. An exceptional case is the sequence with only 1f-right, i.e., 1f-right is executed and all following steps are skipped. The behavior shown in Fig. 8 is less restrictive than that shown in Fig. 7. The reason Figs. 7 and 8 look structurally different may be due to the fact that the input event log is incomplete.

Original languageEnglish (US)
Article number8730480
Pages (from-to)548
Number of pages1
JournalIEEE Transactions on Automation Science and Engineering
Volume17
Issue number1
DOIs
StatePublished - Jan 2020

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Erratum: Process Mining to Discover Shoppers' Pathways at a Fashion Retail Store Using a WiFi-Base Indoor Positioning System(IEEE Transactions on Automation Science and Engineering(2017) 14: 4(1786–1792)Doi:10.1109/TASE.2017.2692961)'. Together they form a unique fingerprint.

Cite this