A Method of Description on the Data Association Based on Granulation Trees

Yan Shuo,Yan Lin 보안공학연구지원센터 2015 International Journal of Database Theory and Appli Vol.8 No.2

To investigate the association of data with other data in reality, the research begins with data sets which are divided into different partitions. Because each partition consists of granules and owns a level, all the partitions constitute a granulation set whose elements are the granules. As a hierarchy system, the granulation set together with the inclusion relation gives rise to a structure called a granulation tree. The research on the data association establishes a method to describe the associations of the data in a granulation tree with the data in another granulation tree. The method involves a necessary and sufficient condition used to check the data associations. Because the necessary and sufficient condition is bound up with the upper approximation, the study also develops a way of investigation into rough sets. As an example, a practical problem is modeled by granulation trees, and the associations of the data in a granulation tree with the data in another granulation tree can be examined by use of the necessary and sufficient condition. Meanwhile, because the study is closely linked to granules and alterations of granularity, the process can be viewed as an approach to research on granular computing

음의 연관성 규칙 생성을 위한 Jaccard 비유사성 측도들의 고찰

류재열,박희창 한국자료분석학회 2013 Journal of the Korean Data Analysis Society Vol.15 No.6

The popularity of smart devices today has led to an exponential growth of data. Therefore, as currently existing way of collecting and handling data came to it’s limit, the need of big data analysis became prominent. A typical technique of big data analysis is data mining, and association rule is largely used among such techniques. Such association rule can be largely divided into positive and negative association rules; negative association rule searches item sets that are in contrary relationship in between items, and also, just as positive association rule, investigates relations using measurements like negative support, negative confidence, and negative lift. From among these, negative support and lift can result in same value when preceding and succeeding items switch places, whereas the value of negative confidence will change, making it difficult creating negative association rule using negative confidence. In this paper, to solve these problems, we studied Jaccard dissimilarity measures used to clustering analysis and multi-dimensional data analysis, and explored their utility as association thresholds for negative association rule. 오늘날 스마트 기기의 급속한 보급으로 데이터의 양이 기하급수적으로 증가하게 되었다. 이로 인해 기존에 사용해오던 방법으로는 데이터를 수집하고 분석하는데 한계가 발생함에 따라 빅 데이터 분석의 필요성이 대두되었다. 빅 데이터 분석의 가장 대표적인 기법이 데이터마이닝이며, 이 기법들 중에서 연관성 규칙이 많이 활용되고 있다. 연관성 규칙을 활용하면 항목들 간의 유용한 규칙을 발견할 수 있으며, 또한 이를 수치화 할 수 있어서 다양한 조직에서 합리적인 의사결정을 위해 이용되고 있다. 이러한 연관성 규칙은 크게 양의 연관성과 음의 연관성 규칙으로 나누어 살펴볼 수 있는데, 이들 중에서 음의 연관성 규칙은 항목들 간에 서로 배반의 관계에 있는 항목집합을 탐색하며, 양의 연관성 규칙과 동일하게 음의 지지도, 음의 신뢰도, 음의 향상도 등의 측도를 사용하여 관계를 규명한다. 이들 중에서 지지도와 향상도의 경우에는 전항과 후항이 바뀌더라도 같은 값을 얻을 수 있지만 신뢰도의 경우에는 값이 달라져서 기존의 음의 신뢰도를 가지고는 음의 연관성 규칙을 생성해내는 데 곤란을 겪을 수 있다. 이러한 문제를 해결하기 위해 본 논문에서는 군집분석이나 다차원 분석에서 활용되는 Jaccard 유사성 측도에 대한 여러 가지 유형의 비유사성 측도를 고찰하였으며, 이들을 음의 연관성 규칙을 평가하기 위한 기준으로 적용 가능성 여부를 탐색하였다.

Relation for the Measure of Association and the Criteria of Association Rule in Ordinal Database

박희창,이호순 한국데이터정보과학회 2005 한국데이터정보과학회지 Vol.16 No.2

One of the well-studied problems in data mining is the search for association rules. Association rules are useful for determining correlations between attributes of a relation and have applications in marketing, financial and retail sectors. There are three criteria of association rule; support, confidence, lift. The goal of association rule mining is to find all the rules with support and confidence exceeding some user specified thresholds. We can know there is association between two items by the criteria of association rules. But we can not know the degree of association between two items. In this paper we examine the relation between the measures of association and the criteria of association rule for ordinal data.

Extraction of Expert Knowledge Based on Hybrid Data Mining Mechanism

김진성(Jin Sung Kim) 한국지능시스템학회 2004 한국지능시스템학회논문지 Vol.14 No.6

This paper presents a hybrid data mining mechanism to extract expert knowledge from historical data and extend expert systems’ reasoning capabilities by using fuzzy neural network (FNN)-based learning & rule extraction algorithm. Our hybrid data mining mechanism is based on association rule extraction mechanism, FNN learning and fuzzy rule extraction algorithm. Most of traditional data mining mechanisms are depended on association rule extraction algorithm. However, the basic association rule-based data mining systems has not the learning ability. Therefore, there is a problem to extend the knowledge base adaptively. In addition, sequential patterns of association rules can’t represent the complicate fuzzy logic in real-world. To resolve these problems, we suggest the hybrid data mining mechanism based on association rule-based data mining, FNN learning and fuzzy rule extraction algorithm. Our hybrid data mining mechanism is consisted of four phases. First, we use general association rule mining mechanism to develop an initial rule base. Then, in the second phase, we adopt the FNN learning algorithm to extract the hidden relationships or patterns embedded in the historical data. Third, after the learning of FNN, the fuzzy rule extraction algorithm will be used to extract the implicit knowledge from the FNN. Fourth, we will combine the association rules (initial rule base) and fuzzy rules. Implementation results show that the hybrid data mining mechanism can reflect both association rule-based knowledge extraction and FNN-based knowledge extension.

Data Outsourcing based on Secure Association Rule Mining Processes

V. Sujatha,Debnath Bhattacharyya,P. Silpa Chaitanya,Tai-hoon Kim 보안공학연구지원센터 2015 International Journal of Security and Its Applicat Vol.9 No.3

Data mining is the process of extracting information from data warehousing applications. Data outsourcing is the major task in present days, for accessing services and other features of the database processing. But sometimes this process may achieve to split among various parties with recommended data items in analyzing of the data. Data security is one of the key processes in outsourcing data to various outside users. Traditionally Fast Distribution Mining algorithm was proposed for securing distributed data. This paper addresses a problem by secure association rules over partitioned data in both horizontal and vertical representation. A secure frequency developed algorithm is used for doing above process efficiently in partitioned data, which includes services of the data in outsourcing process. Frequent item sets are used to access services in outsourcing data in recent application development data mining. Our proposed work maintains efficient security over vertical and horizontal view of representation in secure mining applications. The result shows that algorithm timing is desirable for big size data for security considerations using association rule mining operations in real time application development.

A Study of Association Rule Applicationusing Self-Organizing Map for Fused Data

조광현,박희창 한국데이터정보과학회 2008 한국데이터정보과학회지 Vol.19 No.1

Currently, Gyeongnam province is executing the social index survey to the provincials every year. But, this survey has the limit of the analysis as execution of the different survey per 3 year cycles. The solution for this problem is data fusion technique. Data fusion is the process of combining multiple data in order to provide information of tactical value to the user. Data fusion is generally defined as the use of techniques that collect to combine data including multiple sources in order to raise the quality of information. Data fusion is also called "data combination" or "data matching". Data fusion is divided into five branch types which are exact matching, judgemental matching, probability matching, statistical matching, and data linking. But, data fusion doesn't mean the ultimate result. Therefore, efficient analysis for the fused data is also important. In this study, we suggest the application methodology of association rule using SOM for the fused data of statistical survey data.

교통사고 데이터의 마이닝을 위한 연관규칙 학습기법과 서브그룹 발견기법의 비교

김정민(Jeongmin Kim),류광렬(Kwang Ryel Ryu) 한국지능정보시스템학회 2015 지능정보연구 Vol.21 No.4

Traffic accident is one of the major cause of death worldwide for the last several decades. According to the statistics of world health organization, approximately 1.24 million deaths occurred on the world’s roads in 2010. In order to reduce future traffic accident, multipronged approaches have been adopted including traffic regulations, injury-reducing technologies, driving training program and so on. Records on traffic accidents are generated and maintained for this purpose. To make these records meaningful and effective, it is necessary to analyze relationship between traffic accident and related factors including vehicle design, road design, weather, driver behavior etc. Insight derived from these analysis can be used for accident prevention approaches. Traffic accident data mining is an activity to find useful knowledges about such relationship that is not well-known and user may interested in it. Many studies about mining accident data have been reported over the past two decades. Most of studies mainly focused on predict risk of accident using accident related factors. Supervised learning methods like decision tree, logistic regression, k-nearest neighbor, neural network are used for these prediction. However, derived prediction model from these algorithms are too complex to understand for human itself because the main purpose of these algorithms are prediction, not explanation of the data. Some of studies use unsupervised clustering algorithm to dividing the data into several groups, but derived group itself is still not easy to understand for human, so it is necessary to do some additional analytic works. Rule based learning methods are adequate when we want to derive comprehensive form of knowledge about the target domain. It derives a set of if-then rules that represent relationship between the target feature with other features. Rules are fairly easy for human to understand its meaning therefore it can help provide insight and comprehensible results for human. Association rule learning methods and subgroup discovery methods are representing rule based learning methods for descriptive task. These two algorithms have been used in a wide range of area from transaction analysis, accident data analysis, detection of statistically significant patient risk groups, discovering key person in social communities and so on. We use both the association rule learning method and the subgroup discovery method to discover useful patterns from a traffic accident dataset consisting of many features including profile of driver, location of accident, types of accident, information of vehicle, violation of regulation and so on. The association rule learning method, which is one of the unsupervised learning methods, searches for frequent item sets from the data and translates them into rules. In contrast, the subgroup discovery method is a kind of supervised learning method that discovers rules of user specified concepts satisfying certain degree of generality and unusualness. Depending on what aspect of the data we are focusing our attention to, we may combine different multiple relevant features of interest to make a synthetic target feature, and give it to the rule learning algorithms. After a set of rules is derived, some postprocessing steps are taken to make the ruleset more compact and easier to understand by removing some uninteresting or redundant rules. We conducted a set of experiments of mining our traffic accident data in both unsupervised mode and supervised mode for comparison of these rule based learning algorithms. Experiments with the traffic accident data reveals that the association rule learning, in its pure unsupervised mode, can discover some hidden relationship among the features. Under supervised learning setting with combinatorial target feature, however, the subgroup discovery method finds good rules much more easily than the association rule learning method that requires a lot of efforts to

기준 확인 측도와 확률적 흥미도 측도와의 관계 탐색

박희창 한국자료분석학회 2014 Journal of the Korean Data Analysis Society Vol.16 No.1

빅 데이터는 기존 데이터베이스 관리도구로는 다루기 힘든 대용량의 정형 또는 비정형 데이터를 분석하는 기술을 의미하며, 오늘날 정부와 다양한 민간 분야에서 빅 데이터의 가치에 주목하고 있다. 이러한 빅 데이터의 대표적인 분석 기법이 데이터마이닝이며, 이들 기법 중에서 연관성 규칙이 많이 활용되고 있다. 이는 빅 데이터에 존재하고 있는 항목들 간의 상호 관련성을 찾아내는 기법으로서 항목들 사이의 지지도, 신뢰도, 향상도 등의 흥미도 측도를 기준으로 상호 관련성 여부를 측정한다. 이와는 별개로 흥미도 측도의 관점에서 기준 확인 측도의 성질, 기준 확인 측도에 대한 연관성 평가 기준에 대한 만족 여부, 그리고 기본적인 연관성 평가 측도인 지지도, 신뢰도, 그리고 향상도 등과의 관계가 규명된 바 있다. 그러나 이들 측도들은 모두 양의 값만을 가지기 때문에 연관성의 방향을 파악하기가 곤란하므로 기준 확인 측도와의 비교가 연관성 평가의 관점에서 큰 의미를 부여할 수 없다. 이에 본 논문에서는 비대칭적 기준 확인 측도에 대해 연관성의 방향을 알 수 있는 확률적 흥미도 측도와의 관계를 수식을 통해 유도하는 동시에 예제를 통해 변화하는 양상을 고찰하였다. 그 결과, 확률적 흥미도 측도의 값이 증가할수록 본 논문에서 고려하는 모든 측도들이 증가하는 양상을 보이고 있다. 또한 확률적 흥미도 측도가 양의 값을 가지면 본 논문에서 고려하는 비대칭 기준 확인 측도들도 모두 양의 값을 가지고, 그 값이 음이면 이들 비대칭 기준 확인 측도들도 모두 음의 값을 가지는 것을 확인하였다. Big data is a general term used to describe the huge amount of unstructured and semi-structured data a organization creates. Big data analytics is the process of examining voluminous amounts of data of a variety of types to uncover hidden patterns, unknown correlations and other useful information. Such information can provide competitive advantages over rival organizations and result in business benefits, such as more effective marketing and increased revenue. In most organizations such as government and various companies, they are actively participating in big data techniques development. Data mining technique is the method to find useful information in a big database. The most widely used data mining technique is to generate association rules. In this paper, we found the relationship between asymmetric confirmation measures and probabilistic interestingness measure through the formulas, and compared confirmation measures with probabilistic interestingness measure using some simulation data. As the result, we could distinguish the direction of association rule by confirmation measures, and interpret degree of association operationally by them.

Relation for the Measure of Association and the Criteria of Association Rule in Ordinal Database

Hee Chang Park,Ho Soon Lee 한국데이터정보과학회 2005 한국데이터정보과학회지 Vol.16 No.2

One of the well-studied problems in data mining is the search for association rules. Association rules are useful for determining correlations between attributes of a relation and have appications in marketing, financial and retail sectors. There are three criteria of association rule; support. confidence, lift. "Ire goal of association rule mining is to find all the rules with support and confidence exceeding some user specified thresholds. We can know there is association between two items by the criteria of association rules. But we can not know the degree of association between two items. In this paper we examine the relation between the measures of association and the criteria of association rule for ordinal data.

연관성 규칙 기술에서 순수 교차 엔트로피 측도의 제안

박희창 한국자료분석학회 2018 Journal of the Korean Data Analysis Society Vol.20 No.2

Data mining, one of the technologies that are attracting attention in the big data era, is a technique to search for useful information hidden in a data set as if it is a surplus of beads. Among the data mining techniques, association rule is one of the techniques used in data mining techniques to search for relationship among interested objects based on various interestingness measures (Park, 2017a). For confidence evaluation of the association rule, there are positive confidence, two types of negative confidence, and inverse confidence. But using only positive confidence can cause problem with the purity of the association rule. To solve this problem, we proposed a pure cross entropy measure that considers contrasted confidence in each state instead of the cross entropy measure using marginal probability. As a result of the comparison between the four kinds of simulations, both the cross entropy and the pure cross entropy decreased and then increased. In addition, the value of the pure cross entropy is larger than the cross entropy, so that it can be used more clearly when evaluating association. 빅 데이터 시대에 주목받고 있는 기술 중의 하나인 데이터 마이닝은 구슬을 꿰어야 보배가 되는 것처럼 데이터 집합에 숨겨진 유용한 정보를 탐색하는 기법이다. 데이터마이닝 기법 중에서 연관성 규칙 기술은 다양한 흥미도 측도를 근거로 하여 관심 있는 대상들 간의 연관성 유무를 탐색하는 것으로 데이터마이닝 기법 중에서 많이 활용되고 있는 기법 중의 하나이다(Park, 2017a). 연관성 규칙의 평가 기준 중에서 신뢰도는 양의 신뢰도, 두 종류의 음의 신뢰도, 그리고 역의 신뢰도가 있으나 보통의 경우에는 양의 신뢰도만을 사용함으로써 연관성 규칙의 순수성에 문제가 야기될 수 있다. 이러한 점을 해결하기 위해 본 논문에서는 주변 확률을 이용한 교차 엔트로피 측도 대신 각 상태에서 대립되는 신뢰도를 고려한 순수 교차 엔트로피 측도를 제안하고, 예제를 통해 비교하였다. 네 종류의 모의실험을 통해서 비교해본 결과, 교차 엔트로피와 순수 교차 엔트로피 둘 다 감소하다가 증가하는 것으로 나타났다. 또한 교차 엔트로피와 순수 교차 엔트로피 중에서 후자의 값의 크기가 전자보다 더 커서 연관성을 평가하고자 하는 경우에 좀 더 명확하게 이용할 수 있으며, 값의 변화 폭도 후자가 더 큰 것으로 나타나서 후자의 측도가 더 바람직한 것으로 나타났다.