Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

UPER: Efficient Utility-driven Partially-ordered Episode Rule Mining

Hong Lin, Wensheng Gan, Junyu Ren, Philip S. Yu

TL;DR

The utility of POERs is defined and an algorithm called UPER is proposed to discover high-utility partially-ordered episode rules, intending to discover more valuable rules.

Abstract

Episode mining is a fundamental problem in analyzing a sequence of numerous events. For discovering strong relationships between events in a complex event sequence, episode rule mining is proposed. However, both the episode and episode rules have strict requirements for the order of events. Hence, partially-ordered episode rule mining (POERM) is designed to loosen the constraints on the ordering, i.e., events in the antecedents and consequents of the rule can be unordered, and POERM has been applied to real-life event prediction. In this paper, we consider the utility of POERM, intending to discover more valuable rules. We define the utility of POERs and propose an algorithm called UPER to discover high-utility partially-ordered episode rules. In addition, we adopt a data structure named NoList to store the necessary information, analyze the expansion of POERs in detail, and propose several pruning strategies (namely WEUP, REUCSP, and REEUP) to reduce the number of candidate rules. Finally, we conduct experiments on several datasets to demonstrate the effectivene

UPER: Efficient Utility-driven Partially-ordered Episode Rule Mining

TL;DR

The utility of POERs is defined and an algorithm called UPER is proposed to discover high-utility partially-ordered episode rules, intending to discover more valuable rules.

Abstract

Episode mining is a fundamental problem in analyzing a sequence of numerous events. For discovering strong relationships between events in a complex event sequence, episode rule mining is proposed. However, both the episode and episode rules have strict requirements for the order of events. Hence, partially-ordered episode rule mining (POERM) is designed to loosen the constraints on the ordering, i.e., events in the antecedents and consequents of the rule can be unordered, and POERM has been applied to real-life event prediction. In this paper, we consider the utility of POERM, intending to discover more valuable rules. We define the utility of POERs and propose an algorithm called UPER to discover high-utility partially-ordered episode rules. In addition, we adopt a data structure named NoList to store the necessary information, analyze the expansion of POERs in detail, and propose several pruning strategies (namely WEUP, REUCSP, and REEUP) to reduce the number of candidate rules. Finally, we conduct experiments on several datasets to demonstrate the effectivene
Paper Structure (17 sections, 1 theorem, 10 figures, 8 tables, 3 algorithms)

This paper contains 17 sections, 1 theorem, 10 figures, 8 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Frequent Episode Mining
  4. Utility-driven Episode Mining
  5. Episode Rule Mining
  6. Preliminary and Problem Statement
  7. The UPER Algorithm
  8. Definition and Pruning Strategies
  9. Data Structure
  10. UPER Algorithm
  11. Experiments
  12. Datasets and Experimental Setup
  13. Impact of the Utility Threshold on Runtime
  14. Impact of the Utility Threshold on Memory Cost
  15. Impact of XSpan
  16. ...and 2 more sections

Key Result

Theorem 1

Assuming that $r'$: $X$$\rightarrow$$Y'$ is from the rule $r$: $X$$\rightarrow$$Y$ expanding the event $e$ at $T_i$. According to definition de:search-interval and Fig. fig:expansion, since the Y expansion occurs only in interval I, interval II and interval III, we can get $Y'$.end$\geq$Y.end and $Y

Figures (10)

  • Figure 1: A complex event sequence.
  • Figure 2: The search interval of $r$ expansion chen2021mining.
  • Figure 3: REUCS when XSpan = 1 YSpan = 1, and XYSpan = 3.
  • Figure 4: The NoList of $F$: {$B$, $E$} (left) and $r$: {$B$} $\rightarrow$ {$C$, $E$} (right) when XSpan = 3, YSpan = 2 and XYSpan = 3.
  • Figure 5: The framework of UPER.
  • ...and 5 more figures

Theorems & Definitions (14)

  • Definition 1: Complex event sequence
  • Definition 2: Event set and partially-ordered episode rule fournier2021mining
  • Definition 3: Occurrence and interval
  • Definition 4: Time duration constraint fournier2021mining
  • Definition 5: Overlapping and non-overlapping occurrences
  • Definition 6: Support and confidence of an episode rule fournier2021mining
  • Definition 7: Utility of an event set
  • Definition 8: Utility of an episode rule
  • Definition 9: Rule expansion
  • Definition 10: Search interval
  • ...and 4 more