Search

MotivationAdvances in sequencing technologies have led to sequencing of genomes of a multitude of organisms. However, draft genomes of many of these organisms contain a large number of gaps due to repeats in genomes, low sequencing coverage and limitations in sequencing technologies. Although there exist several tools for filling gaps, many of these do not utilize all information relevant to gap filling.ResultsHere, we present a probabilistic method for filling gaps in draft genome assemblies using second generation reads based on a generative model for sequencing that takes into account information on insert sizes and sequencing errors. Our method is based on the expectation-maximization (EM) algorithm unlike the graph based methods adopted in the literature. Experiments on real biological datasets show that this novel approach can fill up large portions of gaps with small number of errors and misassemblies compared to other state of the art gap filling tools.Availability and ImplementationThe method is implemented using C++ in a software named “Filling Gaps by Iterative Read Distribution (Figbird)”, which is available at: https://github.com/SumitTarafder/Figbird.Contactatif@cse.buet.ac.bdSupplementary informationSupplementary data are available at Bioinformatics online.

Feature selection is one of the most frequent tasks in data mining applications. Its ability to remove useless and redundant features improves the classification performance and gains knowledge about a given problem makes feature selection a common first step in data mining. In many feature selection applications, we need to combine the results of different feature selection processes. The two most common scenarios are the ensembles of feature selectors and the scaling up of feature selection methods using a data division approach. The standard procedure is to store the number of times every feature has been selected as a vote for the feature and then evaluate different selection thresholds with a certain criterion to obtain the final subset of selected features. However, this method is suboptimal as the relationships of the features are not considered in the voting process. Two redundant features may be selected a similar number of times due to the different sets of instances used each time. Thus, a voting scheme would tend to select both of them. In this article, we present a new approach: instead of using only the number of times a feature has been selected, the approach considers how many times the features have been selected together by a feature selection algorithm. The proposal is based on constructing an undirected graph where the vertices are the features, and the edges count the number of times every pair of instances has been selected together. This graph is used to select the best subset of features, avoiding the redundancy introduced by the voting scheme. The proposal improves the results of the standard voting scheme in both ensembles of feature selectors and data division methods for scaling up feature selection.

In recent years, various deep learning architectures have been proposed to solve complex challenges (e.g. spatial dependency, temporal dependency) in traffic domain, which have achieved satisfactory performance. These architectures are composed of multiple deep learning techniques in order to tackle various challenges in traffic tasks. Traditionally, convolution neural networks (CNNs) are utilized to model spatial dependency by decomposing the traffic network as grids. However, many traffic networks are graph-structured in nature. In order to utilize such spatial information fully, it's more appropriate to formulate traffic networks as graphs mathematically. Recently, various novel deep learning techniques have been developed to process graph data, called graph neural networks (GNNs). More and more works combine GNNs with other deep learning techniques to construct an architecture dealing with various challenges in a complex traffic task, where GNNs are responsible for extracting spatial correlations in traffic network. These graph-based architectures have achieved state-of-the-art performance. To provide a comprehensive and clear picture of such emerging trend, this survey carefully examines various graph-based deep learning architectures in many traffic applications. We first give guidelines to formulate a traffic problem based on graph and construct graphs from various kinds of traffic datasets. Then we decompose these graph-based architectures to discuss their shared deep learning techniques, clarifying the utilization of each technique in traffic tasks. What's more, we summarize some common traffic challenges and the corresponding graph-based deep learning solutions to each challenge. Finally, we provide benchmark datasets, open source codes and future research directions in this rapidly growing field., 21pages, 11figures

Vertex classification is an important graph mining technique and has important applications in fields such as social recommendation and e-Commerce recommendation. Existing classification methods fail to make full use of the graph topology to improve the classification performance. To alleviate it, we propose a Dual Graph Wavelet neural Network composed of two identical graph wavelet neural networks sharing network parameters. These two networks are integrated with a semi-supervised loss function and carry out supervised learning and unsupervised learning on two matrixes representing the graph topology extracted from the same graph dataset, respectively. One matrix embeds the local consistency information and the other the global consistency information. To reduce the computational complexity of the convolution operation of the graph wavelet neural network, we design an approximate scheme based on the first type Chebyshev polynomial. Experimental results show that the proposed network significantly outperforms the state-of-the-art approaches for vertex classification on all three benchmark datasets and the proposed approximation scheme is validated for datasets with low vertex average degree when the approximation order is small.

Latches have the advantages of timing-borrowing, smaller cell area, lower input capacitance, and lower power compared to flip-flops (FFs). This paper presents a CAD flow that converts any arbitrarily complex single-clock-domain FF-based RTL design into an efficient 3-phase latch-based design. The flow includes a novel 3-phase aware retiming algorithm for power and area optimization. Recent research has demonstrated significant power benefits of an automatic conversion algorithm from FF-based to 3-phase latch-based designs. This paper proposes a new retiming algorithm for 3-phase latch designs that further reduces the number of required latches while also considering the impact of clock gating. Post place-and-route results demonstrate that our new 3-phase designs achieve retiming algorithm yields 23.5% and 23.9% average power reductions compared to more traditional FF and master-slave based alternatives across a board range of benchmarks with no degradation in performance and on average less area.

Telemedicine is one of the most rapidly developing areas of healthcare and it plays an increasing role in modern medicine. As the amount of data and demand for features increase, the data paths are becoming ever-more complex. Owing to this, it is vital in telemedicine to find a proper balance between consistency and availability under any given circumstances. However, making a trade-off can significantly influence the quality of the data. This study seeks to get an in-depth view of the problem by considering a real-world telemedicine use-case and elaborating the formal system specification of the scenario. After evaluating the specification, the constructed state graph is examined using graph coloring and other graph algorithms.

Background:Mining Twitter streaming posts (i.e. tweets) to find events or the topics of interest has become a hot research problem. In the last decade, researchers have come up with various techniques like bag-of-words techniques, statistical methods, graph-based techniques, topic modelling approaches, NLP and ontology-based approaches, machine learning and deep learning methods for detecting events from tweets. Among these techniques, the graph-based technique is efficient in capturing the latent structural semantics in the tweet content by modelling word co-occurrence relationships as a graph and able to capture the activity dynamics by modelling the user-tweet and user-user interactions.Discussion:This article presents an overview of different event detection techniques and their methodologies. Specifically, this paper focuses on graph-based event detection techniques in Twitter and presents a critical survey on these techniques, their evaluation methodologies and datasets used. Further, some challenges in the area of event detection in Twitter along with future directions of research are presented.Conclusion:A Microblogging services and online social networking sites like Twitter provides a massive amount of valuable information on real-world happenings. There is a need for mining this information, which will help in understanding the social interest and effective decision making on various emergencies. However, event detection techniques need to be efficient in terms of time and memory and accurate for processing such voluminous, noisy and fastarriving information from Twitter.

Complex structural variants (CSVs) are genomic alterations that have more than two breakpoints and are considered as the simultaneous occurrence of simple structural variants. However, detecting the compounded mutational signals of CSVs is challenging through a commonly used model-match strategy. As a result, there has been limited progress for CSV discovery compared with simple structural variants. Here, we systematically analyzed the multi-breakpoint connection feature of CSVs, and proposed Mako, utilizing a bottom-up guided model-free strategy, to detect CSVs from paired-end short-read sequencing. Specifically, we implemented a graph-based pattern growth approach, where the graph depicts potential breakpoint connections, and pattern growth enables CSV detection without pre-defined models. Comprehensive evaluations on both simulated and real datasets revealed that Mako outperformed other algorithms. Notably, validation rates of CSVs on real data based on experimental and computational validations as well as manual inspections are around 70%, where the medians of experimental and computational breakpoint shift are 13 bp and 26 bp, respectively. Moreover, the Mako CSV subgraph effectively characterized the breakpoint connections of a CSV event and uncovered a total of 15 CSV types, including two novel types of adjacent segment swap and tandem dispersed duplication. Further analysis of these CSVs also revealed the impact of sequence homology on the formation of CSVs. Mako is publicly available at https://github.com/xjtu-omics/Mako.

In light of the emergence of Database-as-a-Service paradigm, secure data outsourcing has become one of the crucial challenges which strongly imposes itself. In such a scenario, access control is considered as a major challenge. In fact, access control policies of the data owner must be preserved when data is moved to the cloud. Here, we address this problem by considering inference leakage that could be produced by exploiting functional dependencies. The proposed approach is based on vertical partitioning to produce a set of secure sub-schemas stored in separated partitions in the distributed system. Then, we extend this approach by presenting a secure query processing model to preserve access control policies when querying data from distributed partitions. The effectiveness of our algorithms is confirmed through observations from a variety of conducted experiments.

Over the last decade, due to exponential growth in IoT devices and weak security mechanisms, the IoT is now facing more security challenges than ever before, especially botnet malware. There are many security solutions in detecting botnet malware on IoT devices. However, detecting IoT botnet malware, particularly multi-architecture botnets, is challenging. This paper proposes a graphically structured feature extraction mechanism integrated with reinforcement learning techniques in multi-architecture IoT botnet detection. We then evaluate the proposed approach using a dataset of 22849 samples, including actual IoT botnet malware, and achieve a detection rate of 98.03 with low time consumption. The proposed approach also achieves reliable results in detecting the new IoT botnet (has a new architecture-processor) not appearing in the training dataset at 96.69. To promote future research in the field, we share relevant datasets and source code.

In this article, we consider a finite-elementapproximation of the Bellman equation for the optimal control of switched systems. We show that the problem belongs to a special class that we studied in a previous work, for which we developed an efficient solution algorithm. As an application, we present the problem of generating parking maneuvers for self-driving vehicles on two typical urban parking scenarios. The vehicle is described by four different switched systems in which every switching is associated to a penalization term. In this way, we obtain parking paths that have a small number of direction changes and have a simple structure.

Iterative combinatorial auctions are known to resolve bidder preference elicitation problems. However, winner determination is a known key bottleneck that has prevented widespread adoption of such auctions, and adding a time-bound to winner determination further complicates the mechanism. As a result, heuristic-based methods have enjoyed an increase in applicability. We add to the growing body of work in heuristic-based winner determination by proposing an ant colony metaheuristic–based anytime algorithm that produces optimal or near-optimal winner determination results within specified time. Our proposed algorithm resolves the speed versus accuracy problem and displays superior performance compared with 20 past state-of-the-art heuristics and two exact algorithms, for 94 open test auction instances that display a wide variety in bid-bundle composition. Furthermore, we contribute to the literature in two predominant ways: first, we represent the winner determination problem as one of finding the maximum weighted path on a directed cyclic graph; second, we improve upon existing ant colony heuristic–based exploration methods by implementing randomized pheromone updating and randomized graph pruning. Finally, to aid auction designers, we implement the anytime property of the algorithm, which allows auctioneers to stop the algorithm and return a valid solution to the winner determination problem even if it is interrupted before computation ends.

Positive and Unlabeled Learning (PUL) uses unlabeled documents and a few positive documents for retrieving a set of “interest” documents from a text collection. Usually, PUL approaches are based on the vector space model. However, when dealing with semi-supervised learning for text classification or information retrieval, graph-based approaches have been proved to outperform vector space model-based approaches. So, in this article, a graph-based approach for PUL is proposed: Label Propagation for Positive and Unlabeled Learning (LP-PUL). The proposed framework consists of three steps: (i) building a similarity graph, (ii) identifying reliable negative documents, and (iii) performing label propagation to classify the remaining unlabeled documents as positive or negative. We carried out experiments to measure the impact of the different choices in each step of the proposed framework. We also demonstrated that the proposal surpasses the classification performance of other PUL (RC-SVM, PU-LP, and PE-PUC) or one-class learning ( k -NN-based, k -Means-based, and Dense Autoencoder) algorithms in terms of F 1 . Considering the best results of any algorithm used in the experimental evaluation, PU-PUL can improve the classification performance from 2 % , when using only 1 labeled document, to 28 % , when 30 labeled documents are employed.

Periodic pattern mining in time series database plays an important part in data mining. However, most existing algorithms consider only the count of each item, but do not consider about the value of each item. To consider the value of each item on periodic pattern mining in time series databases, Chanda et al. proposed an algorithm called WPPM. In their algorithm, they construct the suffix trie to store the candidate pattern at first. However, the suffix trie would use too much storage space. In order to decrease the processing time for constructing the data structure, in this paper, we propose two data structures to store the candidates. The first data structure is Weighted Paired Matrix. After scanning the database, we will transform the database into the matrix type, and it is used for the second data structures. Therefore, our algorithm not only can decrease the usage of the memory space, but also the processing time. Because we do not need to use so much time to construct so many nodes and edges. Moreover, wealso consider the case of incremental mining for the increase of the data length. From the performance study, we show that our proposed algorithm based on the Weighted Direction Graphis more efficient than the WPPMalgorithm.

Multi-view semi-supervised classification (MSSC) focuses on exploring information from multiple views of labeled and unlabeled data to boost classification performance. However, most of the existing methods build models for each view individually, therefore, the potential relationship between different views can’t be fully explored. Additionally, they either focus on the correlations for consistency or maximize the independence for complementarity. Consistency and complementarity are equally important for multi-view learning. Therefore, this work proposes a novel Graph-based Remodeling Network for MSSC (GRNet), which can explore the potential relationship of multiple views and balance the consistency and complementation adaptively. Specifically, this model integrates multiple views and then generates reformed pseudo views by an attention-based ensemble learning strategy. Moreover, to exploit the information of unlabeled data, this work introduces graph regularization for such a goal. Extensive experiments on several datasets demonstrate the effectiveness and efficiency of the proposed method.

The aim of this paper is to improve the best known solution of an important problem, the DNA Word Design problem, which has its roots in Bioinformatics and Coding Theory. The problem is to design DNA codes that satisfy some combinatorial constraints. The constraints considered are: minimum Hamming distance, fixed GC content and the reverse complement Hamming distance. The problem is modeled as a maximum independent set problem. Existing complex approaches for the maximum independent set problem, suitable for large graphs, were tested. In order to tackle large instances, libraries for external memory computations and sampling techniques were investigated. Eventually, we succeed in finding good lower bounds for the instances that were analyzed.

In this paper we present a graph-based framework that, utilizing relations between groups of System-calls, detects whether an unknown software sample is malicious or benign, and classifies a malicious software to one of a set of known malware families. In our approach we propose a novel graph representation of dependency graphs by capturing their structural evolution over time constructing sequential graph instances, the so-called Temporal Graphs. The partitions of the temporal evolution of a graph defined by specific time-slots, results to different types of graphs representations based upon the information we capture across the capturing of its evolution. The proposed graph-based framework utilizes the proposed types of temporal graphs computing similarity metrics over various graph characteristics in order to conduct the malware detection and classification procedures. Finally, we evaluate the detection rates and the classification ability of our proposed graph-based framework conducting a series of experiments over a set of known malware samples pre-classified into malware families.

The purpose of this paper was to introduce the concepts of very thin multigroup, nondistributive (very thin) multirings, zero-divisor elements of multirings and zero-divisor graphs based on zero-divisor elements of multirings. In order to realize the article’s goals, we consider the relationship between finite nondistributive (very thin) multirings and multirings and construct nondistributive (very thin) multirings based on given ring. By some conditions on prime numbers, finite (nondistributive) very thin multirings are constructed. Zero-divisor graph based on zero-divisor set of (nondistributive) (very thin) multirings are introduced, so we investigate of some necessity and sufficiency conditions such that compute of order and size of these zero-divisor graphs. Also, the notations of derivable zero-divisor graphs and derivable zero-divisor subgraphs are introduced and is showed that some multipartite graphs are derivable zero-divisor graphs, all complete graphs, and cyclic graphs are derivable zero-divisor subgraphs.

Understanding motion patterns is of great importance to analyze the behavior of objects in the vigilance area. Grouping the motion patterns into clusters in such a way that similar motion patterns lie in same cluster and the inter-cluster variance is maximized in a challenging work. Variation in the duration of trajectory patterns in terms of time or number of points in them (even in the trajectories from same cluster) makes it more difficult to correctly classify in respective clusters while using full length trajectories, local clue can be used along with the global information. Trajectories can be segmented into distinctive parts and local contribution of these parts can be used to improve the performance of the system. In this work, we have formulated the trajectory classification problem into graph based similarity problem using Douglas–Peucker (DP) algorithm, Complete Bipartite Graphs (CBG), and Minimum Spanning Tree (MST). Local behavior of objects has been analyzed using their motion segments and Dynamic Time Warping (DTW) has been used for finding similarity among motion trajectories. Class-wise global and local costs have been computed using DTW, CBG, and MST and their fusion has been done using Particle Swarm Optimization (PSO) to improve the classification rate. Trajectory datasets, namely T15, LabOmni, and CROSS have been used in experiments. The proposed method yields encouraging results and outperforms the state of the art techniques.

Sequential product recommendation, aiming at predicting the products that a target user will interact with soon, has become a hotspot topic. Most of the sequential recommendation models focus on learning from users’ interacted product sequences in a purely data-driven manner. However, they largely overlook the knowledgeable substitutable and complementary relations between products. To address this issue, we propose a novel Substitutable and Complementary Graph-based Sequential Product Recommendation model, namely, SCG-SPRe. The innovations of SCG-SPRe lie in its two main modules: (1) The module of interactive graph neural networks jointly encodes the high-order product correlations in the substitutable graph and the complementary graph into two types of relation-specific product representations. (2) The module of kernel-enhanced transformer networks adaptively fuses multiple temporal kernels to characterize the unique temporal patterns between a candidate product to be recommended and any interacted product in a target behavior sequence. Thanks to the seamless integration of the two modules, SCG-SPRe obtains candidate-dependent user representations for different candidate products to compute the corresponding ranking scores. We conduct extensive experiments on three public datasets, demonstrating SCG-SPRe is superior to competitive sequential recommendation baselines and validating the benefits of explicitly modeling the product-product relations.

Creating a coherent summary of the text is a challenging task in the field of Natural Language Processing (NLP). Various Automatic Text Summarization techniques have been developed for abstractive as well as extractive summarization. This study focuses on extractive summarization which is a process containing selected delineative paragraphs or sentences from the original text and combining these into smaller forms than the document(s) to generate a summary. The methods that have been used for extractive summarization are based on a graph-theoretic approach, machine learning, Latent Semantic Analysis (LSA), neural networks, cluster, and fuzzy logic. In this paper, a semantic graph-based approach SGATS (Semantic Graph-based approach for Automatic Text Summarization) is proposed to generate an extractive summary. The proposed approach constructs a semantic graph of the original Hindi text document by establishing a semantic relationship between sentences of the document using Hindi Wordnet ontology as a background knowledge source. Once the semantic graph is constructed, fourteen different graph theoretical measures are applied to rank the document sentences depending on their semantic scores. The proposed approach is applied to two data sets of different domains of Tourism and Health. The performance of the proposed approach is compared with the state-of-the-art TextRank algorithm and human-annotated summary. The performance of the proposed system is evaluated using widely accepted ROUGE measures. The outcomes exhibit that our proposed system produces better results than TextRank for health domain corpus and comparable results for tourism corpus. Further, correlation coefficient methods are applied to find a correlation between eight different graphical measures and it is observed that most of the graphical measures are highly correlated.

Spatial audio is an important component of realistic simulations in virtual environments. However, there have been shortcomings in existing spatial audio rendering systems, due to the computational cost of rendering realistic sound propagation. In this article, we present a novel spatial audio rendering method using 3-D graphs to simulate sound propagation in real time. We then ask if such a distinction in accuracy makes a difference to task performance, and test this proposition in a 3-D game. We conclude that improvements in spatial audio rendering do increase task performance.

Multi-label learning deals with the problem which each data example can be represented by an instance and associated with a set of labels, i.e., every example can be classified into multiple classes simultaneously. Most of the existing multi-label learning methods are supervised which cannot deal with such application scenarios where manually labeling the data is very expensive and time-consuming while the unlabeled data are very cheap and easy to obtain. This paper proposes an ensemble learning method which integrates multi-label learning and graph-based semi-supervised learning into one framework. The label correlation consistency is introduced to deal with the multi-label learning. The proposed method has been evaluated on five public multi-label datasets by comparing it with state-of-the-art supervised and semi-supervised multi-label methods according to multiple evaluation metrics to confirm its effectiveness. Experimental results show that the proposed method can achieve the comparable performance compared with the state-of-the-art methods. Furthermore, it is more confident on every single predicted label.

Electrocardiogram (ECG) signals have been widely used for disease diagnosis. Besides, the ECG signals can be used for human identification. In this work, a Tietze pattern and neighborhood component analysis (NCA)-based human identification method is proposed. Our model uses two feature generation methods to extract both statistical and textural features. The Tietze graph is considered to create a pattern of the presented local graph structure (LGS). Both statistical and textural feature generations are not enough to present a high-accurate model. Therefore, a multileveled structure must be created. Tunable Q-factor wavelet transform (TQWT) is employed as a decomposer. The generated/extracted features in each level are merged, and the merged features are selected using NCA. The k-nearest neighbors (kNN) classifier is deployed on the chosen features in the classification phase to obtain predicted values. The recommended method was tested on two ECG signal corpora called ECGID and MIT-BIH. The model achieved 99.12% and 99.94% accuracies on the used ECGID and MIT-BIH datasets, respectively.

Research on digital cultural heritage has raised the importance of providing visitors with relevant assistance before and during their visits. With the advent of the social web, the cultural heritage area is affected by the problem of information overload. Indeed, a large number of available resources have emerged coming from the social information systems (SocIS). Therefore, visitors are swamped with enormous choices in their visited cities. SocIS platforms use the features of collaborative tagging, named folksonomy, to commonly contribute to the management of the shared resources. However, collaborative tagging uses uncontrolled vocabulary which semanti- cally weakens the description of resources, consequently decreases their classification, clustering, thereby their recommendation. Therefore, the shared resources have to be pertinently described to ameliorate their recommendations. In this paper, we aim to enhance the cultural heritage visits by suggesting semantically related places that are most likely to interest a visitor. Our proposed approach represents a semantic graph-based recommender system of cultural heritage places through two steps; (1) constructing an emergent semantic description that semantically augments the place and (2) effectively modeling the emerging graphs representing the semantic relatedness of similar cultural heritage places and their related tags. The experimental evaluation shows relevant results attesting the efficiency of the proposed approach.

Forecasting human poses given a sequence of historical pose frames has several important applications, especially in the domain of smart home safety. Recently, computer vision-based human pose forecasting has made a breakthrough using deep learning technology. However, to implement a practical system deployed on an IoT edge environment, there are still two issues to be addressed. First, existing methods on pose forecasting fail to model the coherent structural information of connected human joints and thus cannot achieve satisfactory prediction accuracy, especially for long-term predictions. Second, a general and static pre-trained prediction model may not perform well in the deployment environment due to the visual domain shift problem. In this article, we propose a hybrid cloud-edge system called GPFS to solve those issues. Specifically, we first introduce a novel graph convolutional neural network (GCN)-based sequence-to-sequence learning method, which enhances the sequence encoder by using a graph to represent both the spatial and temporal connections of the human joints in the input frames. The GCN improves the forecasting accuracy by capturing the motion pattern of each joint as well as the correlations among different human joints over time. Second, to address the domain shift issue and protect data privacy, we extend the system to perform online learning on the IoT edge to adapt the cloud trained general model with online collected on-site domain data. Extensive evaluation on Human 3.6M and Penn Action datasets demonstrates the superiority of our proposed system.

In an era of big data and emergence of disrupting mobility technologies, statistical models have been utilized to uncover the influence of significant factors, and machine learning algorithms have ...

Weakly supervised object detection (WSOD), aiming to detect objects with only image-level annotations, has become one of the research hotspots over the past few years. Recently, much effort has been devoted to WSOD for the simple yet effective architecture and remarkable improvements have been achieved. Existing approaches using multiple-instance learning usually pay more attention to the proposals individually, ignoring relation information between proposals. Besides, to obtain pseudo-ground-truth boxes for WSOD, MIL-based methods tend to select the region with the highest confidence score and regard those with small overlap as background category, which leads to mislabeled instances. As a result, these methods suffer from mislabeling instances and lacking relations between proposals, degrading the performance of WSOD. To tackle these issues, this article introduces a multi-peak graph-based model for WSOD. Specifically, we use the instance graph to model the relations between proposals, which reinforces multiple-instance learning process. In addition, a multi-peak discovery strategy is designed to avert mislabeling instances. The proposed model is trained by stochastic gradients decent optimizer using back-propagation in an end-to-end manner. Extensive quantitative and qualitative evaluations on two publicly challenging benchmarks, PASCAL VOC 2007 and PASCAL VOC 2012, demonstrate the superiority and effectiveness of the proposed approach.

Epileptic EEG signal recognition is an important method for epilepsy detection. In essence, epileptic EEG signal recognition is a typical imbalanced classification task. However, traditional machine learning methods used for imbalanced epileptic EEG signal recognition face many challenges: (1) traditional machine learning methods often ignore the imbalance of epileptic EEG signals, which leads to misclassification of positive samples and may cause serious consequences and (2) the existing imbalanced classification methods ignore the interrelationship between samples, resulting in poor classification performance. To overcome these challenges, a graph-based extreme learning machine method (G-ELM) is proposed for imbalanced epileptic EEG signal recognition. The proposed method uses graph theory to construct a relationship graph of samples according to data distribution. Then, a model combining the relationship graph and ELM is constructed; it inherits the rapid learning and good generalization capabilities of ELM and improves the classification performance. Experiments on a real imbalanced epileptic EEG dataset demonstrated the effectiveness and applicability of the proposed method.