Angelina A. Tzacheva

Action rules mining

Action rules assume that attributes in a database are divided into two groups: stable and flexible. In general, an action rule can be constructed from two rules extracted earlier from the same database. Furthermore, we assume that these two rules describe two different decision classes and our goal is to reclassify objects from one of these classes into the other one. Flexible attributes are essential in achieving that goal because they provide a tool for making hints to a user about what changes within some values of flexible attributes are needed for a given group of objects to reclassify them into a new decision class. A new subclass of attributes called semi‐stable attributes is introduced. Semi‐stable attributes are typically a function of time and undergo deterministic changes (e.g., attribute age or height). So, the set of conditional attributes is partitioned into stable, semi‐stable, and flexible. Depending on the semantics of attributes, some semi‐stable attributes can be treated as flexible and the same new action rules can be constructed. These new action rules are usually built to replace some existing action rules whose confidence is too low to be of any interest to a user. The confidence of new action rules is always higher than the confidence of rules they replace. Additionally, the notion of the cost and feasibility of an action rule is introduced in this article. A heuristic strategy for constructing feasible action rules that have high confidence and possibly the lowest cost is proposed.

Breast cancer detection in gadolinium‐enhanced MR images by static region descriptors and neural networks

To automate the diagnosis of malignancy by classifying breast tissues as negative or positive for malignancy in gadolinium‐enhanced dynamic magnetic resonance (MR) images, using static region descriptors and a neural network classifier.

Mining for interesting action rules

In this paper, we give a strategy for constructing all action rules from a given information system and show that action rules constructed by system DEAR, cover only a small part of all action rules. Clearly, we are not interested in all action rules as we are not interested in extracting all possible rules from an information system. Classical strategies like See5, LERS, CART, Rosetta, Weka discover rules whose classification part is either the shortest or close to the shortest. This approach basically rules out all other classification rules unless they are surprising rules. In this paper, we introduce the notion of cost of an action rule and define interesting action rules as rules of the smallest cost. We give a strategy showing how interesting action rules can be generated from action rules discovered by system DEAR.

Constraint Based Action Rule Discovery with Single Classification Rules

Action rules can be seen as an answer to the question: what one can do with results of data mining and knowledge discovery? Some applications include: medical field, e-commerce, market basket analysis, customer satisfaction, and risk analysis. Action rules are logical terms describing knowledge about possible actions associated with objects, which is hidden in a decision system. Classical strategy for discovering them from a database requires prior extraction of classification rules which next are evaluated pair by pair with a goal to suggest an action, based on condition features in order to get a desired effect on a decision feature. An actionable strategy is represented as a term

Association Action Rules and Action Paths Triggered by Meta-actions

r = [(\omega) \wedge (\alpha \rightarrow \beta)] \Rightarrow [\phi \rightarrow \psi]

In Search for Action Rules of the Lowest Cost

, where i¾?, i¾?, β, i¾?, and i¾?are descriptions of objects or events. The term rstates that when the fixed condition i¾?is satisfied and the changeable behavior (i¾?i¾?β) occurs in objects represented as tuples from a database so does the expectation (i¾?i¾?i¾?). With each object a number of actionable strategies can be associated and each one of them may lead to different expectations and the same to different re-classifications of objects. In this paper we will focus on a new strategy of constructing action rules directly from single classification rules instead of pairs of classification rules. It presents a gain on the simplicity of the method of action rules construction, as well as on its time complexity. We present A*-type heuristic strategy for discovering only interesting action rules, which satisfy user-defined constraints such as: feasibility, maximal cost, and minimal confidence. We, therefore, propose a new method for fast discovery of interesting action rules.

Document image matching using a maximal grid approach

Action rules are built from atomic expressions called atomic action terms and they describe possible transitions of objects from one state to another. They involve changes of values within one decision attribute. Association action rule is similar to an action rule but it may refer to changes of values involving several attributes listed in its decision part. Action paths are defined as sequences of association action rules with the assumption that the last rule in a sequence is as action rule. Early research on action rule discovery usually required the extraction of classification rules before constructing any action rule. Newest algorithms discover action rules and association action rules directly from an information system. This paper presents a strategy for generating association action rules and action paths by incorporating the use of meta-actions and influence matrices. Action paths show the cascading effect of meta-actions leading to a desired goal.

Support confidence and utility of action rules triggered by meta-actions

There are two aspects of interestingness of rules that have been studied in data mining literature, objective and subjective measures ([2, 1, 3, 11, 12]). Objective measures are data-driven and domain-independent. Generally, they evaluate the rules based on their quality and similarity between them. Subjective measures, including unexpectedness, novelty [11], and actionability, are user-driven and domain-dependent. A rule is actionable if user can do an action to his/her advantage based on this rule ([2, 1, 3]). Action rules introduced in [7] and investigated further in [8] are constructed from actionable rules. To construct them, authors assume that attributes in a database are divided into two groups: stable and flexible. Flexible attributes provide a tool for making hints to a user what changes within some values of flexible attributes are needed for a given group of objects to re-classify these objects to another decision class. Ras and Gupta (see [10]) proposed how to construct action rules when information system is distributed with autonomous sites. Additionally, the notion of a cost and feasibility of an action rule is introduced in this paper. A heuristic strategy for constructing feasible action rules which have high confidence and possibly the lowest cost is also proposed. Interestingness of such action rules is the highest among actionable rules.

Model-based bone segmentation from digital X-ray images

A new approach for form document representation using the maximal grid of its frameset is presented. Using image processing techniques, a scanned form is transformed into a frameset composed of a number of cells. The maximal grid is the grid that encompasses all the horizontal and vertical lines in the form and can be easily generated from the cell coordinates. The number of cells from the original frameset, included in each of the cells created by the maximal grid, is then calculated. Those numbers are added for each row and column generating an array representation for the frameset. A novel algorithm for similarity matching of document framesets based on their maximal grid representations is introduced. The algorithm is robust to image noise and to line breaks, which makes it applicable to poor quality scanned documents. The matching algorithm renders the similarity between two forms as a value between 0 and 1. Thus, it may be used to rank the forms in a database according to their similarity to a query form. Several experiments were performed in order to demonstrate the accuracy and the efficiency of the proposed approach.

Music information retrieval with temporal features and timbre

Action rules describe possible transitions of objects from one state to another with respect to a distinguished attribute. Early research on action rule discovery usually required the extraction of classification rules before constructing any action rule. Newest algorithms discover action rules directly from a decision system. We employ a pruning step in action rule generation, through the use of meta-actions. They are nodes of higher-level knowledge, linked with atomic action terms, which show changes triggered within classification attributes. In this paper, we propose improved measures for support and confidence of action rules, as well as we introduce a new measure - the notion of utility of action rules. We perform an experiment in medical domain using Mammographic Mass dataset, where action rules suggest possible ways to re-classify breast tumors from malignant to benign severity class. Results show increased support and confidence for the new proposed measures compared to the standard measures.