Frequent Subgraph Research Articles

Distributed frequent subgraph mining on evolving graph using SPARK

Distributed frequent subgraph mining on evolving graph using SPARK

A survey of pattern mining in dynamic graphs

AbstractGraph data is found in numerous domains such as for the analysis of social networks, sensor networks, bioinformatics, industrial systems, and chemistry. Analyzing graphs to identify useful and interesting patterns is an important research area. It helps understanding graphs, and hence support decision making. Since two decades, many graph mining algorithms have been proposed to identify patterns such as frequent subgraphs, paths, cliques, and trees. But most of them assume that graphs are static. This simplifying assumption makes it easy to design algorithms but discard information about how graphs evolve. This article provides a detailed survey of techniques for mining interesting patterns in dynamic graphs, which can serve both as an introduction and as a guide to recent advances and opportunities in this research area. The main tasks related to mining patterns in dynamic graphs are reviewed such as discovering frequent subgraphs, evolution rules, motifs, subgraph sequences, recurrent and triggering patterns, and trend sequences. In addition, an overview of strategies and approaches to solve dynamic graph mining problems is presented, and their advantages and limitations are highlighted. Various extensions are also discussed such as to discover patterns in data streams and big data. Finally, the article mentions several research opportunities.This article is categorized under: Algorithmic Development > Spatial and Temporal Data Mining Algorithmic Development > Association Rules

Mining Top- k pairs of correlated subgraphs in a large network

We investigate the problem of correlated subgraphs mining (CSM) where the goal is to identify pairs of subgraph patterns that frequently co-occur in proximity within a single graph. Correlated subgraph patterns are different from frequent subgraphs due to the flexibility in connections between constituent subgraph instances and thus, existing frequent subgraphs mining algorithms cannot be directly applied for CSM. Moreover, computing the degree of correlation between two patterns requires enumerating and finding distances between every pair of subgraph instances of both patterns - a task that is both memory-intensive as well as computationally demanding. To this end, we propose two holistic best-first exploration algorithms: CSM-E (an exact method) and CSM-A (a more efficient approximate method with near-optimal quality). To further improve efficiency, we propose a top- k pruning strategy, while to reduce memory footprint, we develop a compressed data structure called R eplica, which stores all instances of a subgraph pattern on demand. Our empirical results demonstrate that the proposed algorithms not only mine interesting correlations, but also achieve good scalability over large networks.

PpiGReMLIN: a graph mining based detection of conserved structural arrangements in protein-protein interfaces

BackgroundProtein-protein interactions (PPIs) are fundamental in many biological processes and understanding these interactions is key for a myriad of applications including drug development, peptide design and identification of drug targets. The biological data deluge demands efficient and scalable methods to characterize and understand protein-protein interfaces. In this paper, we present ppiGReMLIN, a graph based strategy to infer interaction patterns in a set of protein-protein complexes. Our method combines an unsupervised learning strategy with frequent subgraph mining in order to detect conserved structural arrangements (patterns) based on the physicochemical properties of atoms on protein interfaces. To assess the ability of ppiGReMLIN to point out relevant conserved substructures on protein-protein interfaces, we compared our results to experimentally determined patterns that are key for protein-protein interactions in 2 datasets of complexes, Serine-protease and BCL-2.ResultsppiGReMLIN was able to detect, in an automatic fashion, conserved structural arrangements that represent highly conserved interactions at the specificity binding pocket of trypsin and trypsin-like proteins from Serine-protease dataset. Also, for the BCL-2 dataset, our method pointed out conserved arrangements that include critical residue interactions within the conserved motif LXXXXD, pivotal to the binding specificity of BH3 domains of pro-apoptotic BCL-2 proteins towards apoptotic suppressors. Quantitatively, ppiGReMLIN was able to find all of the most relevant residues described in literature for our datasets, showing precision of at least 69% up to 100% and recall of 100%.ConclusionsppiGReMLIN was able to find highly conserved structures on the interfaces of protein-protein complexes, with minimum support value of 60%, in datasets of similar proteins. We showed that the patterns automatically detected on protein interfaces by our method are in agreement with interaction patterns described in the literature.

A Novel Distributed Approach for Frequent Subgraphs Mining Across Cloud Computing System (DistFsm)

A Novel Distributed Approach for Frequent Subgraphs Mining Across Cloud Computing System (DistFsm)

Map Reduce Based Optimized Frequent Subgraph Mining Algorithm for Large Graph Database

Distributed System, plays a vital role in Frequent Subgraph Mining (FSM) to extract frequent subgraph from Large Graph database. It help to reduce in memory requirements, computational costs as well as increase in data security by distributing resources across distributed sites, which may be homogeneous or heterogeneous. In this paper, we focus on the problem related complexity of data arises in centralized system by using MapReduce framework. We proposed a MapReduced based Optimized Frequent Subgrph Mining (MOFSM) algorithm in MapReduced framework for large graph database. We also compare our algorithm with existing methods using four real-world standard datasets to verify that better solution with respect to performance and scalability of algorithm. These algorithms are used to extract subgraphs in distributed system which is important in real-world applications, such as computer vision, social network analysis, bio-informatics, financial and transportation network.

Fast and scalable algorithms for mining subgraphs in a single large graph

Mining frequent subgraphs is an important issue in graph mining. It is defined as finding all subgraphs whose occurrences in the dataset are greater than or equal to a given frequency threshold. In recent applications, such as social networks, the underlying graphs are very large. Algorithms for mining frequent subgraphs from a single large graph have been developing rapidly lately. Among all such algorithms, GraMi is considered the state-of-the-art. However, GraMi still consumes a lot of time and memory in the mining of a large graph. In this paper, we propose two effective strategies to optimize the GraMi algorithm, which help to increase performance as well as reduce memory consumption during execution. Firstly, GraMi only lists all frequent subgraphs, without computing the support of each mined subgraph. This is disadvantageous in decision support systems, which require information about the support of all subgraphs. Therefore, we optimize GraMi to compute the support values during the mining process. Secondly, we apply the strategy of sorting all edges in graphs by their frequencies, which means that edges with low frequencies will be mined first, and vice versa. This sorting strategy can reduce the number of possibly infrequent subgraph candidates, especially on large subgraphs that are usually derived from those edges with high frequency. Thirdly, we apply a parallel processing technique, in which each frequent edge is executed simultaneously in a separate thread, and improve our parallel strategy by combination with the sorting strategy. Our experiments were performed on three real datasets and the results showed that the performance, as well as memory requirements, are better than those of the original GraMi algorithm

An Enhancement of Grami Based on Threshold Policy for Pattern Big Graphs

Pattern Mining is the key mechanism to manage large scale data element. Frequent subgraph mining (FSM) considers isomorphism which is a subprocess of pattern mining is a well-studied problem in the data mining. Graphs are considered as a standard structure in many domains such as protein-protein interaction network in biological networks, wired or wireless interconnection networks, web data, etc. FSM is the task of finding all frequent subgraphs from a given database i.e. a single big graph or database of many graphs, whose support is greater than the given threshold value. Many databases consider small graphs for solving complex problems. The classification of graph depends upon the application requirement. A good mining architecture may prevent a lot of memory and time. This paper follows the Grami structure for the analysis of frequent subgraph mining and also introduces the 20% threshold policy for the enhancement of the directed pattern graphs. The constraint satisfaction problem (CSP) has been discussed and analyzed using the Grami approach. The proposed model is compared to Grami on twitter dataset based on the evaluation of time and memory consumed. The proposed algorithm shows an improvement of 3-4 % for both the parameters. The results show that the performance of Grami approach has been improved which shows a 6.6% reduction in time and 21% improvement in memory consumption using the proposed approach.

FERRARI: an efficient framework for visual exploratory subgraph search in graph databases

Exploratory search paradigm assists users who do not have a clear search intent and are unfamiliar with the underlying data space. Query formulation evolves iteratively in this paradigm as a user becomes more familiar with the content. Although exploratory search has received significant attention recently in the context of structured data, scant attention has been paid for graph-structured data. An early effort for building exploratory subgraph search framework on graph databases suffers from efficiency and scalability problems. In this paper, we present a visual exploratory subgraph search framework called ferrari, which embodies two novel index structures called vaccine and advise, to address these limitations. vaccine is an offline, feature-based index that stores rich information related to frequent and infrequent subgraphs in the underlying graph database, and how they can be transformed from one subgraph to another during visual query formulation. advise, on the other hand, is an adaptive, compact, on-the-fly index instantiated during iterative visual formulation/reformulation of a subgraph query for exploratory search and records relevant information to efficiently support its repeated evaluation. Extensive experiments and user study on real-world datasets demonstrate superiority of ferrari to a state-of-the-art visual exploratory subgraph search technique.

Network Traffic Anomaly Detection Algorithm Based on Intuitionistic Fuzzy Time Series Graph Mining

Network traffic anomaly detection is an important technology in cyberspace security. Combining information entropy theory and a variable ordering heuristic intuitionistic fuzzy time series forecasting model, we present a traffic anomaly detection algorithm based on intuitionistic fuzzy time series graph mining. For multi-dimensional attribute entropy of network traffic data, we establish multiple parallel and independent variable ordering heuristic intuitionistic fuzzy time series forecasting models. At each moment, using the multi-dimensional attribute entropy values as vertices, we construct complete graphs using amplitudes of the change in entropy values and edge weights between vertices defined by similarity, and establish an intuitionistic fuzzy time series graph of the traffic data in the time dimension. We perform frequent subgraph mining on the intuitionistic fuzzy time series graph; build the anomaly vectors based on the mining results, and implement adaptive determination for network traffic anomalies by fitting the anomaly vectors. Comparative experiments on universal datasets verify the superior performance of the algorithm.

A novel facial emotion recognition scheme based on graph mining

Recent years have seen an explosion in graph data from a variety of scientific, social and technological fields. From these fields, emotion recognition is an interesting research area because it finds many applications in real life such as in effective social robotics to increase the interactivity of the robot with human, driver safety during driving, pain monitoring during surgery etc. A novel facial emotion recognition based on graph mining has been proposed in this paper to make a paradigm shift in the way of representing the face region, where the face region is represented as a graph of nodes and edges and the gSpan frequent sub-graphs mining algorithm is used to find the frequent sub-structures in the graph database of each emotion. To reduce the number of generated sub-graphs, overlap ratio metric is utilized for this purpose. After encoding the final selected sub-graphs, binary classification is then applied to classify the emotion of the queried input facial image using six levels of classification. Binary cat swarm intelligence is applied within each level of classification to select proper sub-graphs that give the highest accuracy in that level. Different experiments have been conducted using Surrey Audio-Visual Expressed Emotion (SAVEE) database and the final system accuracy was 90.00%. The results show significant accuracy improvements (about 2%) by the proposed system in comparison to current published works in SAVEE database.

An updated dashboard of complete search FSM implementations in centralized graph transaction databases

Frequent subgraph mining algorithms are widely used in various areas for information analysis. As yet, a handful of algorithms have been proposed and defined in the literature. While several experimental studies were reported, these experiments lack critical information which are important for selecting an implementation of an algorithm for a specific case of use. In this paper, we report on experiments that we carried out on available implementations of complete search Frequent Subgraph Mining (FSM) algorithms. These experiments are conducted in order to choose a suitable FSM solution (i.e., implementation). We identified 32 algorithms in the literature, six of them were selected for our experiments, through a filtering process relying on a set of criteria. Thirteen working implementations of these 6 algorithms are experimented. In this paper, we provide details of the experiments in terms of performance metrics and input variation effect. We propose a preliminary selection of the most efficient FSM solutions for end users, based on the most tested centralized graph-transaction datasets of the literature.

Frequent Subgraph Mining for Graph based Tamil Bibliographic Big Data Analytics

Data analysis can be done in more effectively, when they are represented in the form of graphs. Especially, Frequent Subgraph Mining (FSM) is an important technique for extracting similar patterns in the graphs. Normally, things have been assumed as the graph data we are taking will fit in main memory for their processing. But, as the data grow higher and higher, they will not fit in main memory, rather they need a special framework called MapReduce to put them in a distributed fashion and to process. Many Frequent Subgraph Mining (FSM) algorithms are changing their faces to adopt the MapReduce programming paradigm. Using FSM-H, analysis had been performed on various graph based data like molecular structures, viral patterns etc., Even DBLP i.e., the computer science bibliography data have also been analyzed using this pattern extraction technique. Whereas the details of Tamil Journals and Publications are kept hidden and not available widely to do research on them, for getting their insights to improve the number and quality of the journals; and to give some input for the authors interested to work on Tamil research. In this work, we collected the Tamil journals details from the available data sources and extracted essential patterns using Frequent Subgraph Mining Technique. Also, we presented a detailed statistical analytics on certain frequently happening Tamil Journals and Conferences.

Mining weighted subgraphs in a single large graph

Weighted single large graphs are often used to simulate complex systems, and thus mining frequent subgraphs in a weighted large graph is an important issue that has attracted the attention of many researchers. Within the studies on the mining of frequent subgraphs, the Graph Mining (GraMi) algorithm is considered as a state-of-the-art algorithm. However, this algorithm can only be implemented on graph datasets which have node labels and edge labels, without weights, and this means that the importance of all objects in a large graph is the same. In this paper, we propose a frequent subgraph algorithm on a weighted large graph, called Weighted Graph Mining (WeGraMi), which is based on two effective strategies to mining weighted subgraphs. Firstly, we apply a new strategy to calculate the weight of all mined subgraphs, which is based on the weights of the nodes in that subgraph. Secondly, we apply a search space pruning strategy based on the existing weights; if a frequent subgraph cannot satisfy the given weighting threshold, that subgraph will be pruned, which can reduce the processing time and storage space needed. With both directed and undirected graph datasets, our experimental results show that the runtime as well as the memory requirements of our algorithm are significantly better than those of GraMi.

Generalizing Design of Support Measures for Counting Frequent Patterns in Graphs.

Frequent subgraph mining (FSM) from graphs is an active subject in computer science research. One major challenge in FSM is the development of support measures, which are basically functions that map a pattern to its frequency count in a database. Current state-of-the-art in this topic features a hypergraph-based framework for modeling pattern occurrences which unifies the two main flavors of support measures: the overlap-graph based maximum independent set measure (MIS) and minimum image/instance based (MNI) measures. For the purpose of exploring the middle ground between these two groups and guiding the development of new support measures, we present general sufficient conditions for designing new support measures in hypergraph framework, which can be applied to MNI and other support measures that are not included in the overlap graph framework. We utilize the sufficient conditions to generalize MNI and minimum-instance measure (MI) for designing user-defined linear-time measures. Furthermore, we show that a maximum independent subedge set (MISS) measure developed from the sufficient conditions can fill the gap between MIS and MI in computation complexity and support count.

Algorithm for mining approximate frequent subgraphs in a single graph

Algorithm for mining approximate frequent subgraphs in a single graph

BoVSG: bag of visual SubGraphs for remote sensing scene classification

ABSTRACT Remote sensing scene classification is gaining much more interest in the recent few years for many strategic fields such as security, land cover and land use monitoring. Several methods have been proposed in the literature and they can be divided into three main classes based on the features used: handcrafted features, features obtained by unsupervised learning and those obtained from deep learning. Handcrafted features are generally time consuming and suboptimal. Unsupervised learning based features which have been proposed later gave better results but their performances are still limited because they mainly rely on shallow networks and are not able to extract powerful features. Deep learning based features are recently investigated and gave interesting results. But, they cannot be usually used because of the scarcity of labelled remote sensing images and are also computationally expensive. Most importantly, whatever kind of feature is used, the neighbourhood information of them is ignored. In this paper, we propose a novel remote sensing scene representation and classification approach called Bag of Visual SubGraphs (BoVSG). First, each image is segmented into superpixels in order to summarize the image content while retaining relevant information. Then, the superpixels from all images are clustered according to their colour and texture features and a random label is assigned to each cluster that probably corresponds to some material or land cover type. Thus superpixels belonging to the same cluster have the same label. Afterwards, each image is modelled with a graph where nodes correspond to labelled superpixels and edges model spatial neighbourhoods. Finally, each image is represented by a histogram of the most frequent subgraphs corresponding to land cover adjacency patterns. This way, local spatial relations between the nodes are also taken into account. Resultant feature vectors are classified using standard classification algorithms. The proposed approach is tested on three popular datasets and its performance outperforms state-of-the-art methods, including deep learning methods. Besides its accuracy, the proposed approach is computationally much less expensive than deep learning methods.

Mining clique frequent approximate subgraphs from multi-graph collections

Recently, some algorithms have been reported for mining multi-graph Frequent Approximate Subgraphs (FASs). This kind of algorithm has a large applicability in social network analysis, image classification tasks and clustering, among others. However, all of them mine a large number of patterns. For this reason, some algorithms for mining a subset of FASs have been proposed. In this paper, we propose an efficient algorithm for mining a subset of FASs on multi-graph collections. Our proposed algorithm becomes an alternative for reducing the number of mined FASs by computing the clique FASs. It is important to highlight that, to the best of our knowledge; our proposal is the first algorithm for mining clique FASs on multi-graph collections. Our proposal is compared against other reported solutions and evaluated over several synthetic and real-world multi-graphs. In addition, we show the usefulness of the patterns mined by our proposal for image classification.

High performance frequent subgraph mining on transaction datasets: A survey and performance comparison

Graph data mining has been a crucial as well as inevitable area of research. Large amounts of graph data are produced in many areas, such as Bioinformatics, Cheminformatics, Social Networks, etc. Scalable graph data mining methods are getting increasingly popular and necessary due to increased graph complexities. Frequent subgraph mining is one such area where the task is to find overly recurring patterns/subgraphs. To tackle this problem, many main memory-based methods were proposed, which proved to be inefficient as the data size grew exponentially over time. In the past few years, several research groups have attempted to handle the Frequent Subgraph Mining (FSM) problem in multiple ways. Many authors have tried to achieve better performance using Graphic Processing Units (GPUs) which has multi-fold improvement over in-memory while dealing with large datasets. Later, Google's MapReduce model with the Hadoop framework proved to be a major breakthrough in high performance large batch processing. Although MapReduce came with many benefits, its disk I/O and noniterative style model could not help much for FSM domain since subgraph mining process is an iterative approach. In recent years, Spark has emerged to be the De Facto industry standard with its distributed in-memory computing capability. This is a right fit solution for iterative style of programming as well. In this survey, we cover how high-performance computing has helped in improving the performance tremendously in the transactional directed and undirected aspect of graphs and performance comparisons of various FSM techniques are done based on experimental results.

Basic and personalized pattern-based workflow fragments discovery

With an increasing number of scientific workflows accessible on public repositories, the mechanism for discovering and recommending workflow fragments is important to facilitate the reuse and repurposing of legacy workflows when novel workflows are to be constructed, where dependencies among workflow activities, which implies functional patterns with different types and characteristics, are specified in the specification of workflows. Traditional approaches ignore or seldom consider this aspect, which may have certain influence on the quality of personalized recommendation. To address this challenge, this paper proposes a novel workflow fragment discovery mechanism for personalized requirements, where discovery strategies of basic and personalized patterns are presented independently. Specifically, frequent basic subfunctions are discovered from scientific workflows by applying the frequent subgraph mining algorithm. Similar subfunctions are clustered considering their semantic relevance of topics, and clusters with high functional frequency are assumed as basic patterns. Thereafter, the multi-dimensional representation of scientific workflows is constructed to explore workflow relevance. Workflow clustering is conducted, and frequent personalized functions are discovered from clusters with similar workflows and assumed as personalized patterns under respective contents. For a personalized requirement given in terms of a workflow template, target basic patterns and candidate subfunctions are discovered, and they compose the backbone structure of solution in a novel coverage strategy. Candidate personalized patterns are applied to cover remaining functionalities of requirement. An optimal solution is obtained through atomic service optimization. Evaluation results show that this technique is accurate on discovering fragment solutions for personalized requirements in comparison with the state-of-the-art techniques.