Synthetic Instances Research Articles

CALEB: A Conditional Adversarial Learning Framework to enhance bot detection

The high growth of Online Social Networks (OSNs) over the last few years has allowed automated accounts, known as social bots, to gain ground. As highlighted by other researchers, many of these bots have malicious purposes and tend to mimic human behavior, posing high-level security threats on OSN platforms. Moreover, recent studies have shown that social bots evolve over time by reforming and reinventing unforeseen and sophisticated characteristics, making them capable of evading the current machine learning state-of-the-art bot detection systems. This work is motivated by the critical need to establish adaptive bot detection methods in order to proactively capture unseen evolved bots towards healthier OSNs interactions. In contrast with most earlier supervised ML approaches which are limited by the inability to effectively detect new types of bots, this paper proposes CALEB, a robust end-to-end proactive framework based on the Conditional Generative Adversarial Network (CGAN) and its extension, Auxiliary Classifier GAN (AC-GAN), to simulate bot evolution by creating realistic synthetic instances of different bot types. These simulated evolved bots augment existing bot datasets and therefore enhance the detection of emerging generations of bots before they even appear. Furthermore, we show that our augmentation approach overpasses other earlier augmentation techniques which fail at simulating evolving bots. Extensive experimentation on well established public bot datasets, show that our approach offers a performance boost of up to 10% regarding the detection of new unseen bots. Finally, the use of the AC-GAN Discriminator as a bot detector, has outperformed former ML approaches, showcasing the efficiency of our end to end framework.

A Diversity-Based Synthetic Oversampling Using Clustering for Handling Extreme Imbalance

Imbalanced data are typically observed in many real-life classification problems. However, mainstream machine learning algorithms are mostly designed with the underlying assumption of a relatively well-balanced distribution of classes. The mismatch between reality and algorithm assumption results in a deterioration of classification performance. One form of approach to address this problem is through re-sampling methods, although its effectiveness is limited; most re-sampling methods fail to consider the distribution of minority and majority instances and the diversity within synthetically generated data. Diversity becomes increasingly important when minority data becomes more sparse, as each data point becomes more valuable. They should all be considered during the generation process instead of being regarded as noise. In this paper, we propose a cluster-based diversity re-sampling method, combined with NOAH algorithm. Neighbourhood-based Clustering Diversity Over-sampling (NBCDO) is introduced with the aim to complement our previous cluster-based diversity algorithm Density-based Clustering Diversity Over-sampling (DBCDO). It first uses a neighbourhood-based clustering algorithm to consider the distribution of both minority and majority class instances, before applying NOAH algorithm to encourage diversity optimisation during the generation of synthetic instances. We demonstrate the implementation of both cluster-based diversity methods by conducting experiments over 10 real-life datasets with ≤ 5% imbalance ratio and show that our proposed cluster-based diversity algorithm (NBCDO, DBCDO) brings performance improvements over its comparable methods (DB-SMOTE, MAHAKIL, KMEANS-SMOTE, MC-SMOTE).

Learning efficient in-store picking strategies to reduce customer encounters in omnichannel retail

Omnichannel retailers are reinventing stores to meet the growing demand of the online channel. Several retailers now use stores as supporting distribution centers to offer quicker Buy-Online-Pickup-In-Store (BOPS) and Ship-From-Store (SFS) services. They resort to in-store picking to serve online orders using existing assets. However, in-store picking operations require picker carts traveling through store aisles, competing for store space, and possibly harming the offline customer experience. To learn picking policies that acknowledge interactions between pickers and offline customers, we formalize a new problem called Dynamic In-store Picker Routing Problem (diPRP). This problem considers a picker that tries to pick online orders (seeking) while minimizing customer encounters (hiding) – preserving the offline customer experience. We model the problem as a Markov Decision Process (MDP) and solve it using a hybrid solution approach comprising mathematical programming and reinforcement learning components. Computational experiments on synthetic instances suggest that the algorithm converges to efficient policies. We apply our solution approach in the context of a large European retailer to assess the proposed policies regarding the number of orders picked and customers encountered. The learned policies are also tested in six different retail settings, demonstrating the flexibility of the proposed approach. Our work suggests that retailers should be able to scale the in-store picking of online orders without jeopardizing the experience of offline customers. The policies learned using the proposed solution approach reduced the number of customer encounters by up to 50%, compared to policies solely focused on picking orders. Thus, to pursue omnichannel strategies that adequately trade-off operational efficiency and customer experience, retailers cannot rely on actual simplistic picking strategies, such as choosing the shortest possible route.

Iterative minority oversampling and its ensemble for ordinal imbalanced datasets

Ordinal classification of imbalanced datasets is a challenging problem that occurs in many real-world applications. The main challenge is to simultaneously consider the classes ordering and imbalanced distribution. Although the classic synthetic instances oversampling techniques can improve the identification of minority classes, they easily incur the damage of the classes ordering when the synthetic instances fall in non-adjacent classes regions. In this paper, we propose a powerful method for handling the imbalanced problem embedded in the ordinal classification, namely Iterative Minority oversampling technique for imbalanced Ordinal Classification (IMOC). Concretely, we first develop an iterative identification procedure to select the minority instance that is hardest to learn. Then, a weighted oversampling probability distribution that respects the ordinal nature is used to generate synthetic minority instances to balance the skewed distribution. Furthermore, two novel ensemble versions are developed to boost the capability of our proposed IMOC. In order to verify the effectiveness and robustness of our proposed methods, an extensive experimental study is carried out on a large number of datasets from real-world applications. The experimental results supported by proper statistical tests indicate that our proposed methods outperform state-of-the-art algorithms in terms of the most frequently used performance measures.

Synthetic minority oversampling using edited displacement-based [formula omitted]-nearest neighbors

Skewed class proportions in real-world datasets present a challenge for machine learning algorithms, as they have a tendency to correctly categorize the majority class while incorrectly classifying the minority class. Such classification disparities hold significant implications, particularly in predictive scenarios involving minority groups, where misclassifying minority instances could lead to adverse outcomes. To tackle this, class imbalance learning has gained attention, with the Synthetic Minority Oversampling Technique (SMOTE) being a notable approach that addresses class imbalance by generating synthetic instances for the minority class based on their feature space neighbors. Despite its effectiveness and simplicity, SMOTE is known to suffer from a noise propagation issue where noisy and uninformative samples are introduced. While various SMOTE variants, including hybrids with undersampling, have been developed to tackle this problem, identifying noisy samples in complex real-world datasets remains a challenge. To address this, our study introduces a new SMOTE-based hybrid approach called SMOTE-centroid displacement-based k-NN (SMOTE-CDNN). SMOTE-CDNN employs centroid displacement for class prediction, which is more robust against noisy data. After SMOTE is applied, noise instances are detected and removed for clearer decision boundaries if their labels predicted by our centroid displacement-based k-NN algorithm are different from the real ones. While our experiments on 24 imbalance datasets demonstrate the resilience and efficiency of our proposed algorithm, which outperforms state-of-art resampling algorithms with various classification models, we acknowledge the need for further investigation into specific dataset characteristics and classification scenarios to determine the generalizability of our approach.

Variable Neighborhood Search for precedence-constrained tasks optimization on heterogeneous systems

Supercomputing power is one of the fundamental pillars of the digital society, which depends on the accurate scheduling of parallel applications in High-Performance Computing (HPC) centers to minimize computing times. However, precedence-constraint task scheduling is a well-known NP-Hard optimization problem, and no optimal polynomial-time algorithm exists to solve it. Therefore, as accurate as possible to the optimal values, heuristic algorithms are of relevant interest. Our new scheduling proposal name is EFT-GVNS, which stands for Earliest Finish Time - General Variable Neighborhood Search. EFT-GVNS uses a Composite Local Search (CLS), making our proposal more efficient than traditional GVNS. EFT-GVNS accuracy against four high-performance algorithms from the state-of-the-art (EDA, EFT-ILS, GRASP-CPA, MPQGA) and one reference algorithm in the literature (HEFT) is studied. Experimental results over four real-world applications (Fpppp, LIGO, Robot, Sparse) and 14 synthetic instances from the literature show that EFT-GVNS outperforms in terms of the median achieved results, with a global improvement of 37.6%, 27.4%, 17.8%, 6.1%, 2.2% to HEFT, EDA, EFT-ILS, GRASP-CPA, and MPQGA, respectively. EFT-GVNS achieves all the 14 optimal values of the synthetic benchmark.

The Hot Spot Coverage Patrol Problem: Formulations and Solution Approaches

When designing a patrol route, it is often necessary to pay more attention to locations with high crime rates. In this paper, we study a patrol routing problem for a fleet of patrol cars patrolling a region with a high-crime neighborhood (HCN) consisting of multiple hot spots. Considering the disorder and chaos in the HCN, at least one patrol car is required in the HCN at any given time during the patrol. We call this routing problem the hot spot coverage patrol problem (HSCPP). In the HSCPP, the importance of a patrol location is quantified by a prize, and the prize is collected if a patrol car visits the location. Our objective is to maximize the sum of prizes collected by the patrol cars, obeying all operational requirements. We propose mathematical formulations and develop several solution approaches for the HSCPP. The global approach consists of finding the routing solution for all patrol cars with a single integer programming (IP) formulation. The partition approach involves first partitioning the region geographically and solving the routing problem in each subregion with two IP formulations. Next, we strengthen the partition approach by developing a column generation (CG) approach in which the initial columns of the CG approach are the solutions generated from the partition approach. We conduct a detailed computational case study using instances based on real crime data from Montgomery County, Maryland. To further understand the computational tractability of our solution approaches, we also perform a sensitivity analysis using synthetic instances under various scenarios. History: Accepted by Erwin Pesch, Area Editor for Heuristic Search & Approximation Algorithms. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.0192 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2022.0192 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .

A bi-objective model for scheduling green investments in two-stage supply chains

Investing in green technologies to increase sustainability in supply chains has become a common practice for two reasons: the first is directly related to the defense of the environment and people’s health to smooth the emissions of pollutants; the second is the increasing consumer awareness of green products. Despite the higher costs of producing with green technologies and processes, there is also a higher markup on the price of products which rewards the former costs. This study proposes a mathematical model for scheduling green investments over time in a two-stage supply chain to minimize the impact of production on the environment and the economic costs deriving from the investment. The resulting bi-objective model has nonlinear constraints and is solved using a commercial solver. Given its complexity, we propose an upper-bound heuristic and a lower-bound model to reduce the optimality gap attained at a given time limit. Tests on synthetic instances have been conducted, and an example demonstrates the applicability and efficacy of the proposed model.

Data-driven hospitals staff and resources allocation using agent-based simulation and deep reinforcement learning

Hospital staff and resources allocation (HSRA) is a critical challenge in healthcare systems, as it involves balancing the demands of patients, the availability of resources, and the need to provide high-quality health in resource-bounded settings. Traditional approaches to HSRA have relied on manual planning and ad-hoc adjustments, which can be time-consuming and usually lead to sub-optimal outcomes. Recent studies show that machine learning solutions are able to produce better HSRA results compared to manual planning. However, these outcomes usually focused on a single hospital and objective. In this paper, we solve the HSRA task using a novel agent-based simulation with a deep reinforcement learning agent. We used real-world data to generate a wide range of synthetic instances that were used to train the HSRA agent. Our results show that the proposed model is able to achieve better outcomes in terms of patient treatment success and cost-effectiveness compared to previous resource allocation algorithms. We show that different planning horizons obtain similar performance in handling anomalies. In addition, we show a second-order polynomial connection between the patient treatment success and both the hospital’s initial budget and funding over time. These results suggest that our approach has the potential to improve the efficiency and effectiveness of HSRA in healthcare systems.

Minimizing Interference-to-Signal Ratios in Multi-Cell Telecommunication Networks

In contemporary wireless communication networks, base stations are organized into coordinated clusters (called cells) to jointly serve the users. However, such fixed systems are plagued by the so-called cell-edge problem: near the boundaries, the interference between neighboring clusters can result in very poor interference-to-signal power ratios. To achieve a high-quality service, it is an important objective to minimize the sum of these ratios over the cells. The most common approach to solving this minimization problem is arguably the spectral clustering method. In this paper, we propose a new clustering approach, which is deterministic and computationally much less demanding than current methods. Simulating on synthetic instances indicates that our methods typically provide higher quality solutions than earlier methods.

Self-adaptive oversampling method based on the complexity of minority data in imbalanced datasets classification

Learning from imbalanced datasets is a nontrivial task for supervised learning community. Traditional classifiers may have difficulties to learn the concept related to the minority class when addressing imbalanced classification and the issues can become more deteriorated in the presence of other complicated aspects: overlapping, outliers and small disjuncts, etc. In this paper, we propose a self-adaptive oversampling algorithm based on the complexity of minority data for dealing with imbalanced datasets classification problems. In the proposed algorithm, various hyperspheres with different radii determined by imbalance ratio and the distances to the nearest enemy neighbors are firstly generated to cover all minority instances provided that they cannot contain any majority instance. Subsequently, the oversampling process is conducted only within these hyperspheres and thus the generated synthetic minority instances cannot intervene within the majority space, eventually avoiding overlapping issues during achieving between-class balance. In addition, a self-adaptive assignment strategy of oversampling sizes is developed based on the minority data complexity, where the hyperspheres with small radii and few instances in them are provided more chances to be oversampled. The strategy will favor addressing the outliers and small disjuncts issues since the hyperspheres covering the outliers and small disjuncts are usually of small sizes and contain few instances, which makes them have more chances to generate synthetic instances and thus eliminate within-class imbalance due to lack of density. Moreover, since the hyperspheres covering boundary minority instances are relatively small and thus are assigned with larger oversampling sizes, the proposed approach can also strengthen the boundary information of minority class, thus favoring the later learning tasks. The extensive experimental results on various simulated and real-world imbalanced datasets show that the proposed method significantly outperforms other state-of-the-art oversampling ones.

Real-World Pickup and Delivery Problem with Transfers

The pickup and delivery problem with transfers generalizes the classical pickup and delivery problem (PDP) by allowing the vehicles to exchange request loads at designated transfer points. Transfers often lead to substantial reductions in transportation costs, yet they come with a significant burden of additional computational complexity. Even meta-heuristic methods are thus limited to instances of at most lower hundreds of requests leaving the desirable benefits unreachable for larger instances. Our approach bypasses the complexities inherent to current methods by deciding about the transfers apriori and thus reducing the problem to a PDP instance. To make as informed decisions as possible, we analyze a broader set of characteristics that may be used to carry out the apriori decisions. We opt to derive and examine multiple such PDP instances to cover different transfer choices. Our analysis of the derived PDP instances then allows their efficient processing in parallel. The proposed framework addresses a large-scale freight transportation problem with real-world characteristics and transfers where typical instances count over 1,200 requests and 300 vehicles. We show the potential of the proposed framework on both real-world and synthetic instances with up to 1,500 requests. The experiments demonstrate that substantial savings may be achieved within favorable runtimes even for very large instances.

Enhanced Recurrent Neural Network (RNN) For Heart Disease Risk Prediction Using Framingham Datasets

Heart disease is one of the leading causes of death globally, which takes 17.9 million lives each year. The existing heart disease prediction techniques have a gap that does not consider the smoking attributes from the heart disease data. So, the accuracy is based on the limited number of medical data and the deep learning model. The existing deep learning models which use the Recurrent Neural Network (RNN) for heart disease prediction consume more processing and analysing time, mainly due to the delay of data retrieval. This delay causes the prediction process to become slower and leads to a moderate prediction only. The backpropagation of the RNN with an update gate and internal memory to carry the updated data cause a minor data glitch that leads to lower accuracy. Therefore, an efficient heart disease prediction model is very crucial to provide early detection among patients. This research proposes a Heart Disease Risk Prediction Model (HDRPM) with an enhanced RNN to improve the prediction accuracy using Framingham heart disease datasets. The specificity and sensitivity are imposed to improve the quality of the predictions. Sensitivity measure is used for detecting patients with heart disease perfectly and specificity measure is used for detecting patients without the disease perfectly. Besides the accuracy and quality of the prediction problem, the imbalance of minority classes in the dataset occurred in most deep learning prediction fields. This research aims to improve the quality of imbalanced Framingham datasets using Synthetic Minority Over-sampling Technique (SMOTe), which will synthetic instances in a small class to be equalized. The existing RNN model faces vanishing gradients that impede the learning of long data sequences. These gradients that carry information in the RNN cells will become smaller gradually till it minimises the parameter updates and leads to poor learning. For this purpose, the presence of multiple Gated Recurrent Unit (GRU) is used to overcome the vanishing gradients and ensure the hidden layers. The neurons of RNN rapidly cater for the essential information during the training and validation phase of the HDRPM. The integration of multiple GRU with the RNN, operating on the Tensorflow as back-end and Keras as the core for the neural network library has increased the performance of the proposed model. The proposed model provides up to 98.78%, the highest accuracy achieved compared to related previous work, which is a quantum neural network model with 98.57. This HDRPM is expected to significantly contribute to early detection of heart disease patients.

Joint location and cost planning in maximum capture facility location under random utilities

We study a joint facility location and cost planning problem in a competitive market under random utility maximization (RUM) models. The objective is to locate new facilities and make decisions on the costs (or budgets) to spend on the new facilities, aiming to maximize an expected captured customer demand, assuming that customers choose a facility among all available facilities according to a RUM model. We examine two RUM frameworks in the discrete choice literature, namely, the additive and multiplicative RUM. While the former has been widely used in facility location problems, we are the first to explore the latter in the context. We numerically show that the two RUM frameworks can well approximate each other in the context of the cost optimization problem. In addition, we show that, under the additive RUM framework, the resultant cost optimization problem becomes highly non-convex and may have several local optima. In contrast, the use of the multiplicative RUM brings several advantages to the competitive facility location problem. For instance, the cost optimization problem under the multiplicative RUM can be solved efficiently by a general convex optimization solver, or can be reformulated as a conic quadratic program and handled by a conic solver available in some off-the-shelf solvers such as CPLEX or GUROBI. Furthermore, we consider a joint location and cost optimization problem under the multiplicative RUM and propose three approaches to solve the problem, namely, an equivalent conic reformulation, a multi-cut outer-approximation algorithm, and a local search heuristic. We provide numerical experiments based on synthetic instances of various sizes to evaluate the performances of the proposed algorithms in solving the cost optimization and the joint location and cost optimization problems.

NanBDOS: Adaptive and parameter-free borderline oversampling via natural neighbor search for class-imbalance learning

Learning class-imbalance data has become a challenging task in machine learning. Oversampling is an effective way to achieve rebalancing between classes by generating new minority instances. However, most existing oversampling methods exhibit significant ill-posedness due to the involvement of k-nearest neighbors. Recently, several novel techniques have been developed using natural neighbors instead of k-nearest neighbors. But these budding techniques fail to take into account the natural distribution of data, resulting in degraded adaptability of natural neighbors to data characteristics. In this paper, we propose an adaptive and completely parameter-free synthetic oversampling method called NanBDOS (short for borderline oversampling via natural neighbor search). NanBDOS first determines the intrinsic neighbor relationship between instances through a natural neighbor search procedure. Then, those informative borderline instances are identified and used as base instances, i.e., sampling seeds. NanBDOS assigns dynamic sampling weights to the base instances so that the data complexity can be well represented. Finally, new synthetic instances are generated by interpolating between the base instances and their natural neighbors. NanBDOS is experimentally compared with seven baseline methods on twenty-four real-world datasets. The results confirm the effectiveness of the proposed method. Moreover, the statistical analysis also indicates its higher-level Friedman ranking.

Clustering-based improved adaptive synthetic minority oversampling technique for imbalanced data classification

Synthetic Minority Oversampling Technique (SMOTE) and some extensions based on it are popularly used to balance imbalanced data. In this study, we concentrate on solving overfitting of the classification model caused by choosing instances to oversample that increase the occurrence of overlaps with the majority class. Our method called Clustering-based Improved Adaptive Synthetic Minority Oversampling Technique (CI-ASMOTE1) decomposes minority instances into sub-clusters according to their connectivity in the feature space and then selects minority sub-clusters which are relatively close to the decision boundary as the candidate regions to oversample. After application of CI-ASMOTE1, new minority instances are only synthesized within each connected region of the selected sub-clusters. Considering the diversity of the synthetic instances in each selected sub-cluster, CI-ASMOTE2 is put forward to extend CI-ASMOTE1 by keeping all features of those instances in the feature space as different as possible. The experimental evaluation shows that CI-ASMOTE1 and CI-ASMOTE2 improve SMOTE and its extensions, especially in the occurrence of overlaps between the minority instances and the majority instances.

COVID-19 Infodemic in Malaysia: Conceptualizing Fake News for Detection

There is an “Infodemic” of COVID-19 in which there are a lot of rumours and information disorders spreading rapidly, the purpose of the study is to build a predictive model for identifying whether the COVID-19 information in the Malay language in Malaysia is real or fake. Under the study of COVID-19 fake news detection, the synthetic minority oversampling technique (SMOTE) is used to generate synthetic instances of real news in the training set after natural language processing (NLP) and before data modelling because the number of fake news is approximately three times greater than that of real news. Logistic regression, Naïve Bayes, decision trees, support vector machines, random forests, and gradient boosting are employed and compared to determine the most suitable predictive model. In short, the gradient-boosting classifier model has the highest value of accuracy and F1-score.

Combining machine learning and numerical modelling for rockburst prediction

ABSTRACT The application of machine learning in geotechnical problems has grown rapidly. However, an issue in phenomena, such as rockburst, is the limited and imbalanced datasets, which deteriorate the reliability of machine learning algorithms. This research proposes a methodology for overcoming such issues by incorporating both the undersampling and oversampling concepts along with the use of numerical modelling. The core of the methodology is the substitution of clusters containing rockburst instances among similar materials by representative synthetic instances derived from numerical modelling on the basis of the BHP index. The models are tuned and calibrated so that their mechanical properties are aligned with the clusters’ centroids and mimic their nearest neighbours in terms of stresses and rockburst intensities. Subsequently, they are subjected to a variety of stresses to fill all rockburst classes. As a result, a balanced synthetic dataset in terms of material and class participation is formed, which is used to train a Random Forest algorithm to predict rockburst. The results attained indicate the potential of the proposed methodology in enhancing the learning process and thus provide reliable long-term rockburst predictions.

A Strengthened SDP Relaxation for Quadratic Optimization Over the Stiefel Manifold

We study semidefinite programming (SDP) relaxations for the NP-hard problem of globally optimizing a quadratic function over the Stiefel manifold. We introduce a strengthened relaxation based on two recent ideas in the literature: (i) a tailored SDP for objectives with a block-diagonal Hessian; and (ii) the use of the Kronecker matrix product to construct SDP relaxations. Using synthetic instances on five problem classes, we show that, in general, our relaxation significantly strengthens existing relaxations, although at the expense of longer solution times.

HUB-GA: A heuristic for universal lists broadcasting using genetic algorithm

Broadcasting is a fundamental problem in the information dissemination area. In classical broadcasting, a message must be sent from one network member to all other members as rapidly as feasible. Although this problem is NP-hard for arbitrary graphs, it has several applications in various fields. As a result, the universal lists model, which replicates some real-world restrictions like the memory limits of nodes in large networks, is introduced as a branch of this problem in the literature. In the universal lists model, each node is equipped with a fixed list and has to follow the list regardless of the originator. As opposed to various applications for the problem of broadcasting with universal lists, the literature lacks any heuristic or approximation algorithm. In this regard, we suggest HUB-GA: A heuristic for universal lists broadcasting with genetic algorithm, as the first heuristic for this problem. HUB-GA works toward minimizing the universal lists broadcast time of a given graph with the aid of genetic algorithm. We undertake various numerical experiments on frequently used interconnection networks in the literature, graphs with clique-like structures, and synthetic instances with small-world model in order to cover many possibilities of industrial topologies. We also compare our results with state-of-the-art methods for classical broadcasting, which is proved to be the fastest model among all. Nevertheless of the substantial memory reduction in the universal list model compared to the classical model, our algorithm finds the same broadcast time as the classical model in diverse situations.