Earticle

As the use of semantic web based on XML increases in the field of data management, a lot of studies to extract useful information from the data stored in ontology have been tried based on association rule mining. Ontology data is advantageous in that data can be freely expressed because it has a flexible and scalable structure unlike a conventional database having a predefined structure. On the contrary, it is difficult to find frequent patterns in a uniformized analysis method. The goal of this study is to provide a basis for extracting useful knowledge from ontology by searching for frequently occurring subgraph patterns by applying transaction-based graph mining techniques to ontology schema graph data and instance graph data constituting ontology. In order to overcome the structural limitations of the existing ontology mining, the frequent pattern search methodology in this study uses the methodology used in graph mining to apply the frequent pattern in the graph data structure to the ontology by applying iterative node chunking method. Our suggested methodology will play an important role in knowledge extraction.

This paper proposes a new weighted mining frequent pattern based on customer’s RFM(Recency, Frequency, Monetary) score for personalized u-commerce recommendation system under ubiquitous computing. An existing recommendation system using traditional mining has the problem, such as delay of processing speed from a cause of frequent scanning a large data, considering equal weight value of every item, and accuracy as well. In this paper, to solve these problems, it is necessary for us to extract the most frequently purchased data from whole data, to consider the weight/importance of attribute of item in order to forecast frequently changing trends by emphasizing the important items with high purchasability and to improve the accuracy of personalized u-commerce recommendation. To verify improved performance, we make experiments with dataset collected in a cosmetic internet shopping mall.

현존하는 빈발 패턴 마이닝 방법은 대부분 시간 효율성을 목표로 하고, 물리적 메모리 사용에 매우 의존적이다. 하지만 빅데이터 시대가 도래함에 따라 실제 세상의 데이터베이스는 급속도로 증가하고 있으며, 그에 따라 기존의 방법으로 현실적인 거대한 양의 데이터를 마이닝하기에 물리적 메모리 공간이 부족한 실정이다. 이러한 문제를 해결하기 위해, 빈발 패턴 마이닝의 메모리 의존성을 줄이기 위한 보조저장장치 기반의 연구들이 진행되었으나, 메모리 기반의 방법들에 비해 처리 시간이 너무 많이 소비된다는 한계가 있었다. 따라서 확장성을 가지며, 기존의 디스크 기반의 방법들에 비해 시간효율성을 높인 새로운 빈발 패턴 마이닝이 필요하게 되었다. 본 논문에서는 빅데이터로부터 빈도 아이템 집합들을 마이닝하기 위해 메모리와 디스크를 함께 사용하는 스택 기반의 새로운 접근법인 PPFP 알고리즘을 제안하였다. PPFP는 빈발 패턴 마이닝 접근법 중 가장 인기 있고 효율적인 접근법 중 하나인 FP-growth를 기반으로 하고 있다. PPFP 마이닝 방법은 다음과 같이 두 단계로 진행된다. (1) IFP-tree 구축: FP-tree를 생성한 후, 새로운 인덱스 번호 부여 방법으로 FP-tree의 각 노드에 인덱스 번호를 부여하고, 이 인덱스 번호가 부여된 FP-tree(IFP-tree)를 테이블로 변환하여(IFP-table) 디스크에 저장한다. (2) PPFP 알고리즘을 이용한 빈발 패턴 마이닝: 스택 기반의 PUSH-POP 방식으로 패턴을 확장시켜 나가며 빈발 패턴을 마이닝한다. 이러한 방식을 통해 메모리 기반의 방법에 비해 반복적으로 많은 시간이 소모되는 연산에 매우 적은 양의 메모리를 활용하여 확장성과 함께 시간효율성 또한 향상시킬 수 있었다. 그리고 기존의 연구 방법들과 비교 실험을 통해 새로운 알고리즘의 성능을 증명하였다.

Most of existing frequent pattern mining methods address time efficiency and greatly rely on the primary memory. However, in the era of big data, the size of real-world databases to mined is exponentially increasing, and hence the primary memory is not sufficient enough to mine for frequent patterns from large real-world data sets. To solve this problem, there are some researches for frequent pattern mining method based on disk, but the processing time compared to the memory based methods took very time consuming. There are some researches to improve scalability of frequent pattern mining, but their processes are very time consuming compare to the memory based methods. In this paper, we present PPFP as a novel disk-based approach for mining frequent itemset from big data; and hence we reduced the main memory size bottleneck. PPFP algorithm is based on FP-growth method which is one of the most popular and efficient frequent pattern mining approaches. The mining with PPFP consists of two setps. (1) Constructing an IFP-tree: After construct FP-tree, we assign index number for each node in FP-tree with novel index numbering method, and then insert the indexed FP-tree (IFP-tree) into disk as IFP-table. (2) Mining frequent patterns with PPFP: Mine frequent patterns by expending patterns using stack based PUSH-POP method (PPFP method). Through this new approach, by using a very small amount of memory for recursive and time consuming operation in mining process, we improved the scalability and time efficiency of the frequent pattern mining. And the reported test results demonstrate them.

In this paper, in order to identify and recognize attack patterns, we propose a Bayesian classification using frequent patterns. In theory, Bayesian classifiers guarantee the minimum error rate compared to all other classifiers. However, in practice this is not always the case owing to inaccuracies in the unrealistic assumption{ class conditional independence) made for its use. Our method addresses the problem of attribute dependence by discovering frequent patterns. It generates frequent patterns using an efficient FP-growth approach. Since the volume of patterns produced can be large, we propose a pruning technique for selection only interesting patterns. Also, this method estimates the probability of a new case using different product approximations, where each product approximation assumes different independence of the attributes. Our experiments show that the proposed classifier achieves higher accuracy and is more efficient than other classifiers.

Group of genes controls the functioning of a cell by complex interactions. These interacting gene groups are called Gene Regulatory Networks (GRNs). Two previous data mining approaches, clustering and classification have been used to analyze gene expression data. While these mining tools are useful for determining membership of genes by homology, they don't identify the regulatory relationships among genes found in the same class of molecular actions. Furthermore, we need to understand the mechanism of how genes relate and how they regulate one another. In order to detect regulatory relationships among genes from time-series Microarray data, we propose a novel approach using frequent pattern mining and chain rule. In this approach, we propose a method for transforming gene expression data to make suitable for frequent pattern mining, and detect gene expression patterns applying FP-growth algorithm. And then, we construct gene regulatory network from frequent gene patterns using chain rule. Finally, we validated our proposed method by showing that our experimental results are consistent with published results.

대부분의 빈발 패턴은 패턴이 트랜잭션 데이터베이스에 나타나는 support를 패턴 interestingness의 핵심 척도로 다루어 왔으나 패턴의 횟수는 패턴의 completeness가 가지는 정보를 최대치로 가정하고 있다. 그러나 실제적으로는 임의의 패턴 X의 completeness는 트랜잭션에서 서로 다르게 나타나기 마련이다. 따라서 패턴이 가지는 정보의 손실을 줄이기 위해서는 가중치에 의한 support와 completeness에 의한 유용한 패턴 마이닝을 고려하여야 한다. 즉, 높은 completeness율을 갖는 패턴은 더 높은 recall로 이어질 수 있고 높은 빈도수를 갖는 패턴은 보다 높은 정밀도로 이어진다. 본 논문에서는 동적인 항목들의 가중치에 따른 적응된 support와 completeness를 고려하는 WSCFPM 패턴 마이닝 알고리즘을 제안한다. 제안한 방법은 모노톤 또는 반 모노톤 속성이 가중치에 의한 support와 completeness에 영향을 미치지 않기 때문에 탐색과정을 줄일 수 있다. 실험결과를 통하여 제안된 알고리즘이 효과적이며 확장성이 좋은 것임을 보인다.

Most of those studies use frequency, the number of times a pattern appears in a transaction database, as the key measure for pattern interestingness. It prerequisites that any interesting pattern should occupy a maximum portion of the transactions it appears. But in our real world scenarios the completeness of any pattern is more likely to become various in transactions. Hence, we should also consider the problem of finding the qualified patterns with the significant values of the weighted support by completeness in order to reduce the loss of information within any pattern in transaction. In these pattern recommendation applications, patterns with higher completeness may lead to higher recall while patterns with higher completeness may lead to higher recall while patterns with higher frequency lead to higher precision. In this paper, we propose a measure of weighted support and completeness and an algorithm WSCFPM(weigted support and completeness frequent pattern mining). Our algorithm handles the invalidation of the monotone or anti-monotone property which does not hold on completeness. Extensive performance analysis show that our algorithm is very efficient and scalable for word pattern mining.

데이터베이스에서의 빈발 패턴 탐색은 일정 횟수 이상 같이 등장하는 항목들의 집합을 찾는 문제이다. 본 논문은 다수의 2 차원 배열들이 주어졌을 때, 이들 내부에서 빈번히 같이 등장하는 부분 구역들을 찾는 문제를 다룬다. 하지만 기존 빈발 패턴 탐색 기법들은 배열 내 원소들의 위치 관계까지 고려하지는 않기 때문에 본 문제에 바로 적용하기는 어렵다. 따라서 본 논문은 다수의 2 차원 배열 내에서 빈번히 같이 발생하는 부분 구역들을 효율적으로 찾는 기법을 제안한다. 본 논문의 선행 연구에서는 주어진 배열들을 두 번 스캔하여 빈발 부분 구역 집합을 찾는 기법을 제안하였다. 본 논문에서는 이를 개선하여 배열들을 한 번만 스캔하고도 빈발 부분 구역 집합을 찾는 효율적인 기법을 제안한다. 이를 위해 제안 방법은 지금까지 탐색된 부분 구역들에 대한 정보를 메모리에 효율적으로 유지한다. 실험결과 제안방법은 기존 방법에 비해 수행시간을 약 30% 단축함을 보였다.

Data mining is an effective method of the discovery of useful information such as rules and previously unknown patterns existing in large databases. The discovery of association rules is an important data mining problem. We have developed a new parallel mining called Distributed Frequent Pattern Tree (abbreviated by DFPT) algorithm on a distributed shared nothing parallel system to detect association rules. DFPT algorithm is devised for parallel execution of the FP-growth algorithm. It needs only two full disk data scanning of the database by eliminating the need for generating the candidate items. We have achieved good workload balancing throughout the mining process by distributing the work equally to all processors. We implemented the algorithm on a PC cluster system, and observed that the algorithm outperformed the Improved Count Distribution scheme.

As the use of semantic web based on XML increases in the field of data management, a lot of studies to extract useful information from the data stored in ontology have been tried based on association rule mining. Ontology data is advantageous in that data can be freely expressed because it has a flexible and scalable structure unlike a conventional database having a predefined structure. On the contrary, it is difficult to find frequent patterns in a uniformized analysis method. The goal of this study is to provide a basis for extracting useful knowledge from ontology by searching for frequently occurring subgraph patterns by applying transaction-based graph mining techniques to ontology schema graph data and instance graph data constituting ontology. In order to overcome the structural limitations of the existing ontology mining, the frequent pattern search methodology in this study uses the methodology used in graph mining to apply the frequent pattern in the graph data structure to the ontology by applying iterative node chunking method. Our suggested methodology will play an important role in knowledge extraction.