Mining Actionable Patterns by Role Models

Abstract

Data mining promises to discover valid and potentially useful patterns in data. Often, discovered patterns are not useful to the user. "Actionability" addresses this problem in that a pattern is deemed actionable if the user can act upon it in her favor. We introduce the notion of "action" as a domain-independent way to model the domain knowledge. Given a data set about actionable features and an utility measure, a pattern is actionable if it summarizes a population that can be acted upon towards a more promising population observed with a higher utility. We present several pruning strategies taking into account the actionability requirement to reduce the search space, and algorithms for mining all actionable patterns as well as mining the top k actionable patterns. We evaluate the usefulness of patterns and the focus of search on a real-world application domain.

Selecting Actionable Patterns from Positive and Negative Sequential Patterns

“…Some methods discuss how to select actionable knowledge [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. actionable knowledge discovery framework was formally defined in [1], which includes concepts, processes and operable deliverables.…”

“…in [11], a new framework was used to measure the statistical dependencies of sequences. In [12], a domainindependent way was applied to model the domain knowledge and several methods were proposed to mine actionable patterns as well as the top k actionable patterns. Actionable association rules were extracted from multiple datasets in [13].…”

“…N the literature lots of efforts have been made to discover actionable knowledge [1][2][3][4][5][6][7][8][9][10][11][12][13][14], to measure the interestingness in association rule mining [15][16][17][18] and to mine Positive Sequential Patterns (PSP) [19,20]. PSP are frequent subsequences which are uncovered from sequence database.…”

Select actionable positive or negative sequential patterns

“…K. Wang, Y. Jiang and A. Tuzhilin introduced the notion of "action" as a domainindependent method to model the domain knowledge. This presents several pruning strategies to reduce the search space and algorithms for mining all actionable patterns, as well as mining the top k actionable patterns [7]. P. Kanikar and D.K.…”

Action Rules Approach to Solving Diagnostic Problems in Clinical Medicine

“…System ARAS is an exception -it follows rule-based approach, but because it does not require to construct pairs of rules in its discovery process, the number of action rules it discovers is comparable to numbers obtained when we follow object-based approaches. Meta-actions [22] are defined as actions which trigger changes of flexible attributes either directly or indirectly because of correlations among certain attributes in the system. Links between meta-actions and changes they trigger within the values of flexible attributes can be defined by a mapping linking meta-actions with changes of attribute values used to describe objects in the decision system.…”