Measures Of Interest Research Articles

Selected group-theoretic aspects of confirmation measure symmetries

The paper considers symmetry properties of rule interestingness measures (in particular: measures of confirmation). Many authors have studied various symmetries, however the discussion on which sets of symmetries should be taken into account, let alone why the particular symmetries are desirable or not, has still not produced a generally recognized consensus. Furthermore, the results published so far neglect the fact that symmetries can be the subject of group theory-based considerations. This paper aims at solving those problems by introducing group-theoretic interpretations of symmetries, indicating that all symmetries can be treated as permutations, and compositions of symmetries as compositions of permutations. Such an interpretation allows us to apply the well-known group-theoretic results to symmetries. In particular, using this approach, we reveal the phenomenon of incompleteness occurring in sets of symmetries considered by different authors, and propose an effective way of controlling it. Moreover, we demonstrate that assessing the symmetries as either desirable or undesirable, as introduced by these authors, brings in inconsistencies, which become evident under the permutation-based interpretation. Finally, we present group-theoretic guidelines to the design of such symmetry sets that are free of the incompleteness and inconsistency phenomena but remain meaningful in the context of rule evaluation.

P-N-RMiner: a generic framework for mining interesting structured relational patterns

Methods for local pattern mining are fragmented along two dimensions: the pattern syntax, and the data types on which they are applicable. Pattern syntaxes include subgroups, n-sets, itemsets, and many more; common data types include binary, categorical, and real-valued. Recent research on relational pattern mining has shown how the aforementioned pattern syntaxes can be unified in a single framework. However, a unified model to deal with various data types is lacking, certainly for more complexly structured types such as real numbers, time of day—which is circular—, geographical location, terms from a taxonomy, etc. We introduce P-N-RMiner, a generic tool for mining interesting local patterns in (relational) data with structured attributes. We show how to handle the attribute structures in a generic manner, by modelling them as partial orders. We also derive an information-theoretic subjective interestingness measure for such patterns and present an algorithm to efficiently enumerate the patterns. We find that (1) P-N-RMiner finds patterns that are substantially more informative, (2) the new interestingness measure cannot be approximated using existing methods, and (3) we can leverage the partial orders to speed up enumeration.

Approximate Functional Dependencies Mining Using Association Rules Specificity Interestingness Measure

Mining Approximate Functional Dependencies (AFDs) from a database may produce valuable interesting relationships among its variables that would be beneficial in several domain applications such as marketing, financial data analysis, biological data analysi s, and intrusion detection. Mining of association rules is concerned with extracting new knowledge from dat abases in the form of patterns and associations among data items. The mining of AFDs still posing special se t of challenges such as the space and time requirements in addition to the quality of the discovered AFD s. In this paper, an approach for AFDs mining is being developed through employing the specificity interestingness measure and its monotonic property used in some association rules mining algorithms. Thi s approach has been tested on a set of test bed of datasets. The results showed an improvement in time requirements and in the number of mined AFDs.

Subjective interestingness of subgraph patterns

The utility of a dense subgraph in gaining a better understanding of a graph has been formalised in numerous ways, each striking a different balance between approximating actual interestingness and computational efficiency. A difficulty in making this trade-off is that, while computational cost of an algorithm is relatively well-defined, a pattern's interestingness is fundamentally subjective. This means that this latter aspect is often treated only informally or neglected, and instead some form of density is used as a proxy. We resolve this difficulty by formalising what makes a dense subgraph pattern interesting to a given user. Unsurprisingly, the resulting measure is dependent on the prior beliefs of the user about the graph. For concreteness, in this paper we consider two cases: one case where the user only has a belief about the overall density of the graph, and another case where the user has prior beliefs about the degrees of the vertices. Furthermore, we illustrate how the resulting interestingness measure is different from previous proposals. We also propose effective exact and approximate algorithms for mining the most interesting dense subgraph according to the proposed measure. Usefully, the proposed interestingness measure and approach lend themselves well to iterative dense subgraph discovery. Contrary to most existing approaches, our method naturally allows subsequently found patterns to be overlapping. The empirical evaluation highlights the properties of the new interestingness measure given different prior belief sets, and our approach's ability to find interesting subgraphs that other methods are unable to find.

Performance Evaluation of Distributed Association Rule Mining Algorithms

Association Rule Mining (ARM) is a popular and well researched method for discovering interesting relations between variables in large databases. It is intended to identify strong rules discovered in databases using different measures of interestingness. Most ARM algorithms focus on a sequential or centralized environment where no external communication is required. Distributed ARM algorithms (DARM), aim to generate rules from different data sets spread over various geographical sites; hence, they require external communications throughout the entire process. DARM algorithm efficiency is highly dependent on data distribution. The Classical algorithms used in DARM are Count Distribution Algorithm (CDA), Fast Distributed Mining (FDM) Algorithm and Optimized Distributed Association Mining (ODAM) Algorithm. This paper presents the implementation details and experimental results of above mentioned algorithms. The paper also highlights the issues of message exchange size in a distributed environment of current DARM algorithms that can affect the communication costs in a distributed environment.

Sentiment Analysis Using Twitter Data in the Context of Natural Calamity

This study is intended to take advantage of the data from social media to get some insights on the situations generated during a natural calamity. Specifically, the case in point is on Twitter during November–December 2015 south Indian floods resulted from heavy rainfall triggered by the annual northeast monsoon in Chennai, the metropolitan. In all, about 2,500 tweets were examined for the sentiment analysis. The approaches used in this work are (i) coarse-grained classification of sentiments into positive and negative as well as fine grained classification of negative sentiments, (ii) finding temporal distribution of sentiments during and after the tenor of calamity and (iii) finding the significance of conjoint expression of different sentiments in single tweets through measure of interestingness. The results of our analysis showed that negative sentiments will be predominant during the incident and a few days after the incident and gradual subsidence with an increase in positive sentiments as an outcome of good relief work, voluntary services and philanthropic support. We also found that among the combination of two sentiments in a single tweet, sadness and disgust combination was more prevalent followed by sadness and anger.

Ontology Knowledge Mining Based Association Rules Ranking

Medical association rules induction is used to discover useful correlations between pertinent concepts from large medical databases. Nevertheless, ARs algorithms produce huge amount of delivered rules and do not guarantee the usefulness and interestingness of the generated knowledge. To overcome this drawback, we propose an ontology based interestingness measure for ARs ranking. According to domain expert, the goal of the use of ARs is to discover implicit relationships between items of different categories such as ‘clinical features and disorders’,‘clinical features and radiological observations’, etc. That's to say, the itemsets which are composed of “similar” items are uninteresting. Therefore, the dissimilarity between the rule's items can be used to judge the interestingness of association rules; the more different are the items, the more interesting the rule is. In this paper, we design a distinct approach for ranking semantically interesting association rules involving the use of an ontology knowledge mining approach. The basic idea is to organize the ontology's concepts into a hierarchical structure of conceptual clusters of targeted subjects, where each cluster encapsulates “similar” concepts suggesting a specific category of the domain knowledge. The interestingness of association rules is, then, defined as the dissimilarity between corresponding clusters. That's to say, the further are the clusters of the items in the AR, the more interesting the rule is. We apply the method in our domain of interest – mammographic domain-using an existing mammographic ontology called Mammo*, with the goal of deriving interesting rules from past experiences, to discover implicit relationships between concepts modeling the domain.*http://sourceforge.net/p/gimimammography/code/HEAD/tree/trunk/owl

A Novel Mapreduce Lift Association Rule Mining Algorithm (Mrlar) for Big Data

Big Data mining is an analytic process used to discover the hidden knowledge and patterns from a massive, complex, and multi-dimensional dataset. Single-processor's memory and CPU resources are very limited, which makes the algorithm performance ineffective. Recently, there has been renewed interest in using association rule mining (ARM) in Big Data to uncover relationships between what seems to be unrelated. However, the traditional discovery ARM techniques are unable to handle this huge amount of data. Therefore, there is a vital need to scalable and parallel strategies for ARM based on Big Data approaches. This paper develops a novel MapReduce framework for an association rule algorithm based on Lift interestingness measurement (MRLAR) which can handle massive datasets with a large number of nodes. The experimental result shows the effi-ciency of the proposed algorithm to measure the correlations between itemsets through integrating the uses of MapReduce and LIM instead of depending on confidence.

A high utility itemset mining algorithm based on subsume index

High utility itemset mining addresses the limitations of frequent itemset mining by introducing measures of interestingness that reflect the significance of an itemset beyond its frequency of occurrence. Among such algorithms, level-wise candidate generation-and-test approaches suffer from the drawbacks of having an immense candidate pool and requiring several database scans. Meanwhile, methods based on pattern growth tend to consume large amounts of memory to store conditional trees. We propose an efficient algorithm, called Index High Utility Itemsets Mine (IHUI-Mine), for application to high utility itemsets. The subsume index, which has been employed to mine frequent itemsets, is extended in IHUI-Mine to the discovery of high utility itemsets. In addition to the enumeration and search strategies inherited from the subsume index, we introduce a new property to specifically accelerate the computation of transaction-weighted utilization for high utility itemsets. Furthermore, given that bitmaps are used for database representation, the real utility of candidates can be verified from the recorded transactions rather than by resorting to the entire database. The computational complexity of IHUI-Mine is analyzed, and tests conducted on publicly available synthetic and real datasets further demonstrate that the proposed algorithm outperforms existing state-of-the-art algorithms.

Fast exhaustive subgroup discovery with numerical target concepts

Subgroup discovery is a key data mining method that aims at identifying descriptions of subsets of the data that show an interesting distribution with respect to a pre-defined target concept. For practical applications the integration of numerical data is crucial. Therefore, a wide variety of interestingness measures has been proposed in literature that use a numerical attribute as the target concept. However, efficient mining in this setting is still an open issue. In this paper, we present novel techniques for fast exhaustive subgroup discovery with a numerical target concept. We initially survey previously proposed measures in this setting. Then, we explore options for pruning the search space using optimistic estimate bounds. Specifically, we introduce novel bounds in closed form and ordering-based bounds as a new technique to derive estimates for several types of interestingness measures with no previously known bounds. In addition, we investigate efficient data structures, namely adapted FP-trees and bitset-based data representations, and discuss their interdependencies to interestingness measures and pruning schemes. The presented techniques are incorporated into two novel algorithms. Finally, the benefits of the proposed pruning bounds and algorithms are assessed and compared in an extensive experimental evaluation on 24 publicly available datasets. The novel algorithms reduce runtimes consistently by more than one order of magnitude.

Discovering market basket patterns using hierarchical association rules

Association rules are a data mining method for discovering patterns of frequent item sets, such as products in a store that are frequently purchased at the same time by a customer (market basket analysis). A number of interestingness measures for association rules have been developed to date, but research has shown that there a dominant measure does not exist. Authors have mostly used objective measures, whereas subjective measures have rarely been investigated. This paper aims to combine objective measures such as support, confidence and lift with a subjective approach based on human expert selection in order to extract interesting rules from a real dataset collected from a large Croatian retail chain. Hierarchical association rules were used to enhance the efficiency of the extraction rule. The results show that rules that are more interesting were extracted using the hierarchical method, and that a hybrid approach of combining objective and subjective measures succeeds in extracting certain unexpected and actionable rules. The research can be useful for retail and marketing managers in planning marketing strategies, as well as for researchers investigating this field.

An optimized fuzzy method for finding actions

Most of machine learning and data mining techniques only focus on finding patterns satisfying technical interestingness measures. They do not consider business interestingness measures in the mining process, so most of mined patterns may not be usable in corresponding domain. Moreover, the techniques do not interpret the patterns in order that business people know what actions to take to support their business decisions. They require a considerable amount of extra work by domain experts to interpret the patterns. Actionable Knowledge Discovery techniques deliver technical and business significant patterns and support automatically converting patterns into deliverables in business-friendly and operable forms such as actions. In this paper, we incorporate fuzzy set theory and attribute flexibility for the enhancement fuzzy cost-effective action mining algorithm (F-CEAMA). We present an algorithm that suggests actions regarding to their practical applicability. The algorithm takes into account attribute flexibility and fuzzy cost of actions and attempts to maximize the fuzzy net profit. The contribution of the work is in taking the output from fuzzy decision tree, using flexibility of its attributes, and producing novel and actionable knowledge through automatic fuzzy post processing. The performance of proposed algorithm is compared to F-CEAMA using several real-life datasets taken from the UCI Machine Learning Repository. Experimental results show that the proposed algorithm outperforms F-CEAMA not only in finding more actions but also in finding actions with more fuzzy net profit.

Discovering in Formative Knowledge using Combined Mining Approach

mining applications involve complex data like multiple heterogeneous data sources, different user preference and create decision making activities. The comprehensive, useful information may not be obtained by using the single data mining method in the form of informative patterns as that would consume more time and space. Combined mining is a hybrid mining approach for mining informative patterns from single or multiple data sources, many features abstraction and merging multiple methods as per the requirements. Some concepts show hybrid or combined mining approach. In this paper, multi method combined mining methodology is designed. Proposed method combine apriori algorithm with Multi-Objective Evolutionary Algorithm. It helps to improve searching for the exact products from complex data. Proposed method combine apriori algorithm with multi-objective evolutionary algorithm to improve searching of documents from complex data. It addressed challenging problems in combined mining and summarized and proposed effective pattern merging and interaction paradigms, combined pattern types, such as pair patterns and cluster patterns, interestingness measures. In the proposed method, by using apriori algorithm, calculate the support and confidence of the frequent item set which improve results of searching by a using multi-objective evolutionary algorithm. Keywordsmining approach, Multi-objective Evolutionary algorithm, Apriori algorithm, Knowledge Discovery combined mining, complex data, data mining, multiple source data mining.

Smart drill-down

We present a data exploration system equipped with smart drill-down, a novel operator for interactively exploring a relational table to discover and summarize "interesting" groups of tuples. Each such group of tuples is represented by a rule. For instance, the rule (a, b, ★, 1000) tells us that there are a thousand tuples with value a in the first column and b in the second column (and any value in the third column). Smart drill-down presents an analyst with a list of rules that together describe interesting aspects of the table. The analyst can tailor the definition of interesting, and can interactively apply smart drill-down on an existing rule to explore that part of the table. In the demonstration, conference attendees will be able to use the data exploration system equipped with smart drill-down, and will be able to contrast smart drill-down to traditional drill-down, for various interestingness measures, and resource constraints.

Interestingness measure for mining sequential patterns in sports

The increasing availabilities of tracking devices, including mobile devices and sports trackers with heart-rate monitors, accelerometers and GPS receivers, have increased the interest in developing fitness applications. The aims of these applications

Hierarchical interestingness measures for association rules with generalization on both antecedent and consequent sides

Pairwise generalized association rules mined from raw data can be used to connect the concepts of multiple ontologies. In this case the items of rules are hierarchically organized and one can use the relations between them in order to reduce rule redundancy. Recently proposed hierarchical interestingness measures address this issue, taking hierarchical information on the antecedent side into account. In this paper, we extend them to the case of considering two hierarchies on both the antecedent and the consequent sides of a rule. The extended measures are then compared with their counterparts as well as with conventional ones. Three real world datasets from the text mining domain with predefined ground truth sets of associations are used for comparison within the framework of instance-based ontology mapping.

Analysing the Quality of Association Rules by Computing an Interestingness Measures

Objective: Association rule mining is one of the data mining process for discovering frequent item set between transaction databases. The main objective of this research work is statistically analyses the quality rules in the apriori algorithm of association rule mining. Methods: An Interestingness measures is a subset of statistical method and it can give the solution for splitting interesting rules within huge association rules. Currently, it has shown around hundred and above measures. Specifically, this study is to concentrate on eight measures such as lift, chi-square, hyper-lift, hyper-confidence, conviction, coverage, leverage and cosine. In this analysis is performed in two places of real databases whereas Agriculture and Medical domain. Findings: At the experimental results, the proposed system is rectified that the problem many interesting rules are eliminated in satisfying the threshold value of support and confidence. Therefore, the user do not confirm that the strength of interest rules may be least by setting the low threshold value. The comparison and correlation measures also obtained along with the interesting rules. There are some measures outperformed than other and thus measures can mostly correlate with the order lift, chi-squared, hyper-lift, hyper-confidence and conviction. The performance of this work is consistently checking in difference size of transaction databases in addition to we identify the unresolved problem of apriori algorithm. Conclusion: Finally, this research concludes that statistical interestingness measures are really helpful for finding interesting rules among large association rules.

Guest editor’s introduction: special issue on quality issues, measures of interestingness and evaluation of data mining models

There are many data mining algorithms and methodologies for various fields and various problematic. Each data mining researcher/practitioner should describe the intrinsic quality of the discovered models and patterns. In addition he/she is faced with assessing the performance of his/her own solution(s) in order to make –fair– comparisons with state of the art approaches. Which methodology, which benchmarks, which measures of performance, which tools, which measures of interest, which scenarios, etc., should be used, and why? Every one should answer these questions, and assessing the quality and the performance of a data mining process is a critical issue. Assessing the quality and the performance is a critical issue for classical situations and even more in the Big Data era where we can not manage this issue with classical and current approaches. Clearly, large scale data, complex data and streaming data bring new challenges (e.g. large scale inference, fake correlations, necessity of perpetual validation). In this setting, the Quality Issues, Measures of Interestingness and Evaluation of data mining models Workshop series provide an open forum for intensive discussions and the exchange of new ideas and best practices on quality issues among data mining researchers. QIMIE focuses on the theory, the techniques, the links between the discovery stage and the quality assessment stage, and the practices that can ensure that the discovered knowledge of a data mining process is of quality. It is organized with the IEEE Task Force on Evaluation and Quality issues (Data Mining and Big Data Analytics Technical Committee, IEEE Computational Intelligence Society).

Mining unexpected patterns using decision trees and interestingness measures: a case study of endometriosis

Because clinical research is carried out in complex environments, prior domain knowledge, constraints, and expert knowledge can enhance the capabilities and performance of data mining. In this paper we propose an unexpected pattern mining model that uses decision trees to compare recovery rates of two different treatments, and to find patterns that contrast with the prior knowledge of domain users. In the proposed model we define interestingness measures to determine whether the patterns found are interesting to the domain. By applying the concept of domain-driven data mining, we repeatedly utilize decision trees and interestingness measures in a closed-loop, in-depth mining process to find unexpected and interesting patterns. We use retrospective data from transvaginal ultrasound-guided aspirations to show that the proposed model can successfully compare different treatments using a decision tree, which is a new usage of that tool. We believe that unexpected, interesting patterns may provide clinical researchers with different perspectives for future research.

Can interestingness measures be usefully visualized?

Abstract The paper presents visualization techniques for interestingness measures. The process of measure visualization provides useful insights into different domain areas of the visualized measures and thus effectively assists their comprehension and selection for different knowledge discovery tasks. Assuming a common domain form of the visualized measures, a set of contingency tables, which consists of all possible tables having the same total number of observations, is constructed. These originally four-dimensional data may be effectively represented in three dimensions using a tetrahedron-based barycentric coordinate system. At the same time, an additional, scalar function of the data (referred to as the operational function, e.g., any interestingness measure) may be rendered using colour. Throughout the paper a particular group of interestingness measures, known as confirmation measures, is used to demonstrate the capabilities of the visualization techniques. They cover a wide spectrum of possibilities, ranging from the determination of specific values (extremes, zeros, etc.) of a single measure, to the localization of pre-defined regions of interest, e.g., such domain areas for which two or more measures do not differ at all or differ the most.