High Imbalance Research Articles

Using a hybrid content-based and behaviour-based featuring approach in a parallel environment to detect fake reviews

Abstract The financial impact of positive reviews has prompted some fraudulent sellers to generate fake product reviews for either promoting their products or discrediting competing products. Many e-commerce portals have implemented measures to detect such fake reviews, and these measures require excellent detectors to be effective. In this work, we propose 133 unique features from the combination of content and behaviour-based features to detect fake reviews using machine learning classifiers. Preliminary results show that these features can provide good results for all datasets tested. Detailed analysis of the results, however, reveals the existence of class imbalance issues for two of the bigger datasets - there is a high imbalance between the accuracies of different classes (e.g., 7.73% for the fake class and 99.3% for the genuine class using a Multilayer Perceptron classifier). We therefore introduce two sampling methods that can improve the accuracy of the fake review class on balanced datasets. The accuracies can be improved to a maximum of 89% for both random under and over-sampling on Convolutional Neural Networks. Additionally, we propose a parallel cross-validation method that can speed up the validation process in a parallel environment.

Hierarchical Microbial Functions Prediction by Graph Aggregated Embedding.

Matching 16S rRNA gene sequencing data to a metabolic reference database is a meaningful way to predict the metabolic function of bacteria and archaea, bringing greater insight to the working of the microbial community. However, some operational taxonomy units (OTUs) cannot be functionally profiled, especially for microbial communities from non-human samples cultured in defective media. Therefore, we herein report the development of Hierarchical micrObial functions Prediction by graph aggregated Embedding (HOPE), which utilizes co-occurring patterns and nucleotide sequences to predict microbial functions. HOPE integrates topological structures of microbial co-occurrence networks with k-mer compositions of OTU sequences and embeds them into a lower-dimensional continuous latent space, while maximally preserving topological relationships among OTUs. The high imbalance among KEGG Orthology (KO) functions of microbes is recognized in our framework that usually yields poor performance. A hierarchical multitask learning module is used in HOPE to alleviate the challenge brought by the long-tailed distribution among classes. To test the performance of HOPE, we compare it with HOPE-one, HOPE-seq, and GraphSAGE, respectively, in three microbial metagenomic 16s rRNA sequencing datasets, including abalone gut, human gut, and gut of Penaeus monodon. Experiments demonstrate that HOPE outperforms baselines on almost all indexes in all experiments. Furthermore, HOPE reveals significant generalization ability. HOPE's basic idea is suitable for other related scenarios, such as the prediction of gene function based on gene co-expression networks. The source code of HOPE is freely available at https://github.com/adrift00/HOPE.

Resampling imbalanced data to detect fake reviews using machine learning classifiers and textual-based features

Fraudulent online sellers often collude with reviewers to garner fake reviews for their products. This act undermines the trust of buyers in product reviews, and potentially reduces the effectiveness of online markets. Being able to accurately detect fake reviews is, therefore, critical. In this study, we investigate several preprocessing and textual-based featuring methods along with machine learning classifiers, including single and ensemble models, to build a fake review detection system. Given the nature of product review data, where the number of fake reviews is far less than that of genuine reviews, we look into the results of each class in detail in addition to the overall results. We recognise from our preliminary analysis that, owing to imbalanced data, there is a high imbalance between the accuracies for different classes (e.g., 1.3% for the fake review class and 99.7% for the genuine review class), despite the overall accuracy looking promising (around 89.7%). We propose two dynamic random sampling techniques that are possible for textual-based featuring methods to solve this class imbalance problem. Our results indicate that both sampling techniques can improve the accuracy of the fake review class—for balanced datasets, the accuracies can be improved to a maximum of 84.5% and 75.6% for random under and over-sampling, respectively. However, the accuracies for genuine reviews decrease to 75% and 58.8% for random under and over-sampling, respectively. We also discover that, for smaller datasets, the Adaptive Boosting ensemble model outperforms other single classifiers; whereas, for larger datasets, the performance improvement from ensemble models is insignificant compared to the best results obtained by single classifiers.

An Ensemble of Deep Learning-Based Multi-Model for ECG Heartbeats Arrhythmia Classification

An automatic system for heart arrhythmia classification can perform a substantial role in managing and treating cardiovascular diseases. In this paper, a deep learning-based multi-model system is proposed for the classification of electrocardiogram (ECG) signals. Two different deep learning bagging models are introduced to classify heartbeats into different arrhythmias types. The first model (CNN-LSTM) is based on a combination of a convolutional neural network (CNN) and long short-term memory (LSTM) network to capture local features and temporal dynamics in the ECG data. The second model (RRHOS-LSTM) integrates some classical features, i.e. RR intervals and higher-order statistics (HOS), with LSTM model to effectively highlight abnormality heartbeats classes. We create a bagging model from the CNN-LSTM and RRHOS-LSTM networks by training each model on a different sub-sampling dataset to handle the high imbalance distribution of arrhythmias classes in the ECG data. Each model is also trained using a weighted loss function to provide high weight for not sufficiently represented classes. These models are then combined using a meta-classifier to form a strong coherent model. The meta-classifier is a feedforward fully connected neural network that takes the different predictions of bagging models as an input and combines them into a final prediction. The result of the meta-classifier is then verified by another CNN-LSTM model to decrease the false positive of the overall system. The experimental results are acquired by evaluating the proposed method on ECG data from the MIT-BIH arrhythmia database. The proposed method achieves an overall accuracy of 95.81% in the “subject-oriented” patient independent evaluation scheme. The averages of F1 score and positive predictive value are higher than all other methods by more than 3% and 8% respectively. The experimental results show the superiority of the proposed method for ECG heartbeats classification compared to many state-of-the-art methods.

ENSEMBLE META CLASSIFIER WITH SAMPLING AND FEATURE SELECTION FOR DATA WITH IMBALANCE MULTICLASS PROBLEM

Ensemble learning by combining several single classifiers or another ensemble classifier is one of the procedures to solve the imbalance problem in multiclass data. However, this approach still faces the question of how the ensemble methods obtain their higher performance. In this paper, an investigation was carried out on the design of the meta classifier ensemble with sampling and feature selection for multiclass imbalanced data. The specific objectives were: 1) to improve the ensemble classifier through data-level approach (sampling and feature selection); 2) to perform experiments on sampling, feature selection, and ensemble classifier model; and 3 ) to evaluate t he performance of the ensemble classifier. To fulfil the objectives, a preliminary data collection of Malaysian plants’ leaf images was prepared and experimented, and the results were compared. The ensemble design was also tested with three other high imbalance ratio benchmark data. It was found that the design using sampling, feature selection, and ensemble classifier method via AdaboostM1 with random forest (also an ensemble classifier) provided improved performance throughout the investigation. The result of this study is important to the on-going problem of multiclass imbalance where specific structure and its performance can be improved in terms of processing time and accuracy.

Feature Selection and Ensemble Learning Techniques in One-Class Classifiers: An Empirical Study of Two-Class Imbalanced Datasets

Class imbalance learning is an important research problem in data mining and machine learning. Most solutions including data levels, algorithm levels, and cost sensitive approaches are derived using multi-class classifiers, depending on the number of classes to be classified. One-class classification (OCC) techniques, in contrast, have been widely used for anomaly or outlier detection where only normal or positive class training data are available. In this study, we treat every two-class imbalanced dataset as an anomaly detection problem, which contains a larger number of data in the majority class, i.e. normal or positive class, and a very small number of data in the minority class. The research objectives of this paper are to understand the performance of OCC classifiers and examine the level of performance improvement when feature selection is considered for pre-processing the training data in the majority class and ensemble learning is employed to combine multiple OCC classifiers. Based on 55 datasets with different ranges of class imbalance ratios and one-class support vector machine, isolation forest, and local outlier factor as the representative OCC classifiers, we found that the OCC classifiers are good at high imbalance ratio datasets, outperforming the C4.5 baseline. In most cases, though, performing feature selection does not improve the performance of the OCC classifiers in most. However, many homogeneous and heterogeneous OCC classifier ensembles do outperform the single OCC classifiers, with some specific combinations of multiple OCC classifiers, both with and without feature selection, performing similar to or better than the baseline combination of SMOTE and C4.5.

Anode Effect Prediction in Hall-Héroult Cells Using Time Series Characteristics

In aluminium production, anode effects occur when the alumina content in the bath is so low that normal fused salt electrolysis cannot be maintained. This is followed by a rapid increase of pot voltage from about 4.3 V to values in the range from 10 to 80 V. As a result of a local depletion of oxide ions, the cryolite decomposes and forms climate-relevant perfluorocarbon (PFC) gases. The high pot voltage also causes a high energy input, which dissipates as heat. In order to ensure energy-efficient and climate-friendly operation, it is important to predict anode effects in advance so that they can be prevented by prophylactic actions like alumina feeding or beam downward movements. In this paper a classification model is trained with aggregated time series data from TRIMET Aluminium SE Essen (TAE) that is able to predict anode effects at least 1 min in advance. Due to a high imbalance in the class distribution of normal state and labeled anode effect state as well as possible model’s weaknesses the final F1 score of 32.4% is comparatively low. Nevertheless, the prediction provides an indication of possible anode effects and the process control system may react on it. Consequent practical implications will be discussed.

Optimal valuation of wind energy projects co-located with battery storage

The electricity market in the UK has been reformed over recent years resulting in the introduction of ‘Contracts for Difference’ as an option for wind producers. They can use these contracts only if at the same time they enter into other market contracts, i.e., Power Purchase Agreements or participation in the wholesale market. Most wind producers prefer the option of holding a long-term Power Purchase Agreement with another party. However, these agreements do not always offer the best terms for a wind producer due to them selling energy at lower electricity prices than those of the wholesale market spot prices. In addition, direct participation in the wholesale market can lead to high imbalance costs for the wind producers. In this paper, we propose an optimal dispatch model solved via dynamic programming optimisation to examine these two different business cases for wind operators alongside the existence (or not) of subsidised Contracts for Difference. A case study with an onshore UK wind farm and a lithium-based battery is provided. The proposed methodology was designed flexibly, so that it can be applied to different wind farm and battery configurations, as well as energy market structures. It is found that it is not always more profitable for the wind operator to hold a Power Purchase Agreement with another party than owning a battery storage unit and trading wind energy directly into the wholesale market. The discount rate of the Power Purchase Agreements, the capital cost of the wind farm and the strike price of the Contracts for Difference are critical factors affecting the profitability of both business cases. In addition, if subsidies are totally removed from both of them, then these are no longer economically justifiable, unless significant reductions are achieved in the capital costs of wind farms and battery storage units. Lastly, we show that considering only the calendar life of the battery (and not the cycle life) leads to significant underestimation of the investment costs.

Clay behavior following macroaggregate breakdown in Ferralsols

An increase in organic matter content is one of the most important factors favoring stabilization of macroaggregates in soils. However, organic matter is also known to inhibit the flocculation of suspended clay, which implies that flocculation is unlikely to happen after clay is released by the breakdown of water-stable macroaggregates. The aim of this study was to investigate whether clay flocculation is possible following the mechanical breakdown of water-stable macroaggregates in Ferralsols. Profiles of three soils weathered from igneous rocks were studied to create a natural gradient of organic matter, macroaggregation, clay dispersion, and flocculation. Macroaggregates stabilized mainly by inorganic bonds showed high water re-flocculable clay content following breakdown. Organic bonds increased the proportion of water-stable macroaggregates, but, following breakdown, the released clay showed a high imbalance in charges, thereby inhibiting its flocculation. Flocculation can still prevail over dispersion in clay saturated with organic matter when the base saturation is low and Al3+ concentration is high, but the charges in surface horizons are too imbalanced to allow flocculation. Highly stable microaggregates (pseudo-silt and pseudo-sand) are favored under both very low and very high organic matter contents.

Automated Clinical Target Volume Delineation for Non-Small Cell Lung Cancer Patients Using Deep 3D Networks

Patient-specific clinical target volume (CTV) delineation is the most difficult step in defining the planning target volume from the gross tumor volume. Delineating patient specific CTVs is a labor-intensive process, prone to human error, with poor inter-observer consistency. Automated volumetric expansions have been used to reduce these concerns, but these ignore patient specific anatomic and clinical variables, as well as the clinical experience of the radiation oncologist. We developed an automated CTV delineation system for lung cancer, which was trained to learn the physician’s CTV contour pattern using a deep 3D neural network. Our network took both patients’ CT images and their corresponding internal target volume (ITV) as input, and output CTV delineation. ITVs were contoured by physicians after studying the internal motion of the target from the 4D CT scans. In our study, the network was trained from twenty-five NSCLC patients in right upper lung lobe (RUL). Then, the trained network was tested on four additional datasets from RUL NSCLC patients. Each CTV delineation output by the trained network was compared with the corresponding CTV contoured by physicians. One challenge to train a 3D network for CTV delineation is the high imbalance between the CTV volume and the background volume. Instead of using hyperparameters to balance as existing methods did, we directly optimize a loss function robust to high imbalance. This improves the stability of the model, especially when the balance dynamically changes in images. Three out of four CTV contours that were automatically delineated by our trained model agree very well with the physician’s CTV contour. Only one has visible discrepancies near the apex of the upper lobe, but still looks reasonable. As shown in Table 1, the performance of the trained neural network was also quantitatively evaluated using precision and recall between the predicted CTVs and the ground truth CTVs. The average precision is 0.902, and recall is 0.871. The recall and precision are the fraction of true volume among the total ground-truth and the fraction of true volume among the predicted volume, respectively. The automated CTV delineation was utilized for RUL NSCLC patients using the deep learning method solely based on CT images with manually contoured ITV. On the limited testing samples, the model contoured CTVs reasonably well. This approach would potentially combine the ease of automated volumetric expansions while incorporating the patient specific anatomic and clinical variables along with the clinical expertise of the radiation oncologist.Tabled 1Abstract 2782; Table; A quantitative evaluation of our automated CTV delineation system on four testing patients.Test #1Test #2Test #3Test #4Recall0.9420.7270.9090.907Precision0.8150.8890.9550.950 Open table in a new tab

Coordinated Bidding Strategy of Wind Farms and Power-to-Gas Facilities Using a Cooperative Game Approach

Power-to-gas (P2G) technology is an effective way to provide additional flexibility to electricity systems with high penetration of renewable energy. Effective P2G bidding strategies in electricity markets can bring in arbitrage opportunities. This article proposes a coordinated bidding strategy for wind farms and P2G facilities using a cooperative game approach. A coordination framework for optimizing the profit of wind farms and P2G facilities is presented, in which wind power participates in day-ahead, real-time and reserve markets, and the P2G facilities participate in day-ahead and reserve markets for arbitrage. Wind farms strategically purchase reserve to avoid high imbalance penalties. As a coalition, wind farms can provide energy to P2G facilities and receive reserve capacity in return. The coordinated bidding strategy is modeled on stochastic programming to maximize the expected joint profit of wind farms and P2G facilities. Nucleolus and Shapley-value based methods are proposed to allocate profits to each player in the coalition. Numerical results based on real world market data demonstrate that the proposed bidding strategy can increase the profits of both wind farms and P2G facilities. Strategically purchasing reserve can reduce the imbalance penalties of wind farms, and participating in reserve markets of P2G facilities can increase profit.

Critical risk determination method of energy-flow network for urban electricity system under extreme heat wave impact

The extreme high temperature weather event caused by global warming has increased the risk of urban electric power supply. Critical risk determination method is useful to the system risk prevention and adaptive strategy formulation. Based on analyzing the failure mechanism of power supply under heat wave impact, the paper put forward the method framework of energy-flow network on electric power system risk evaluation. There are three parts in the framework including energy-flow chart analysis of power supply and consumption, establishment and network analysis of Gephi visualization network, and matrix diagram evaluation of node importance. The visualized network of power system with different years and electricity input from external area were established and compared in the case study of Beijing in China. The analysis results showed that the network structure stability of the power system decreased slightly, which was due to the node energy flow becoming high imbalance such as power input from external area, local power generation and energy consumption. For risk prevention, power infrastructures should focus on the high-voltage transmission and distribution facilities and 500 kV hub substations. Meanwhile, to electric power generation and supply, natural gas supply and power supply from Inner Mongolia Autonomous Region played the key roles. And power consumption should focus on the tertiary industry and living consumption.

Least squares KNN-based weighted multiclass twin SVM

K-nearest neighbor (KNN) based weighted multi-class twin support vector machines (KWMTSVM) is a novel multi-class classification method. In this paper, we propose a novel least squares version of KWMTSVM called LS-KWMTSVM by replacing the inequality constraints with equality constraints and minimized the slack variables using squares of 2-norm instead of conventional 1-norm. This simple modification leads to a very fast algorithm with much better results. The modified primal problems in the proposed LS-KWMTSVM solves only two systems of linear equations whereas two quadratic programming problems (QPPs) need to solve in KWMTSVM. The proposed LS-KWMTSVM, same as KWMTSVM, employed the weight matrix in the objective function to exploit the local information of the training samples. To exploit the inter class information, we use weight vectors in the constraints of the proposed LS-KWMTSVM. If any component of vectors is zero then the corresponding constraint is redundant and thus we can avoid it. Elimination of redundant constraints and solving a system of linear equations instead of QPPs makes the proposed LS-KWMTSVM more robust and faster than KWMTSVM. The proposed LS-KWMTSVM, commensurate as the KWMTSVM, all the training data points into a “1-versus-1-versus-rest” structure, and thus our LS-KWMTSVM generate ternary output {-1,0,+1} which helps to deal with imbalance datasets. Numerical experiments on several UCI and KEEL imbalance datasets(with high imbalance ratio) clearly indicate that the proposed LS-KWMTSVM has better classification accuracy compared with other baseline methods but with remarkably less computational time.

Exploring Connections among Ecosystem Services Supply, Demand and Human Well-Being in a Mountain-Basin System, China.

Stakeholder perception and supply–demand relations are the main challenges and future directions for research on ecosystem services (ES). Based on spatial data and statistical data, we mapped eight key ES supply between 2005–2015 in the Huailai mountain-basin area. Using data from 507 survey questionnaires, we identified the ES demand and the public perceptions of the changes in ES. In addition, we also reveal the characteristics of the spatial distribution of ES demand, analyze the spatial supply–demand matching of ES, and explore the relationships between ES supply–demand and human well-being. From 2005 to 2015, a general upward trend was observed in the supply of four types of product provisioning services, which is different from the trend perceived by the stakeholders. An increasing trend was observed for carbon sequestration and forest recreation, which was in line with the perceptions of change. A spatial mismatch existed between the ES demand and supply, whereby the supply of carbon sequestration, soil conservation, habitat quality, and forest recreation services exceeded the demand in townships in the mountainous and hilly regions. On the other hand, townships located in the valley plains experienced a high imbalance between the demand and the supply. For the four types of product provisioning services, most towns and villages showed a balance in demand and supply. Linking ES supply–demand and human well-being showed that a threshold may exist in the supply–demand imbalance of regulating and supporting services before it begins to impact human well-being. Our study would enrich the theory and methodology research on relationships between ecosystem services and human well-being, and support knowledge to land allocation and management.

Hybrid neural network with cost-sensitive support vector machine for class-imbalanced multimodal data

Although deep learning exhibits advantages in various applications involving multimodal data, it cannot effectively solve the class-imbalance problem. Herein, we propose a hybrid neural network with a cost-sensitive support vector machine (hybrid NN-CSSVM) for class-imbalanced multimodal data. We used a fused multiple-network structure obtained by extracting the features of different modality data, and used cost-sensitive support vector machines (SVMs) as a classifier. To alleviate the insufficiency of learning from minority-class data, our proposed cost-sensitive SVM loss function reflects different weights of misclassification errors from both majority and minority classes, by controlling cost parameters. Additionally, we present a theoretical setting of the cost parameters in our model. The proposed model is validated on real datasets that range from low to high imbalance ratios. By exploiting the complementary advantages of two architectures, the hybrid NN-CSSVM performs excellently, even with data having a minor-class proportion of only 2%.

Bilateral Opportunism in Buyer–supplier Partnerships

ABSTRACT Purpose Literature analyzes interorganizational relationships mostly from a relational perspective by emphasizing partnerships’ positive aspects. Yet business reality is often less romantic and more opportunistic. Especially within the automotive industry opportunistic behaviors often cause extreme reactions in both partners, resulting in poisoned relationships. This study examines the effects of bilateral opportunism between buyers and suppliers on both partners’ relationship induced financial performance, using an automotive industrial setting of new product development (NPD) partnerships. Furthermore, we evaluate the “opportunistic fit” between both partners to assess whether partnerships are characterized by opportunistic parity or unilateral domination. Methodology The empirical analyses are based on primary data from 147 NPD partnerships between tier-1 suppliers and automotive manufacturers (OEMs), exclusively drawn from the suppliers’ perspective. Data were statistically analyzed using structural equation modeling (SEM) with SPSS AMOS. Findings Our findings suggest that large parts of NPD partnerships between suppliers and OEMs are opportunistically lopsided, meaning that they are unilaterally dominated. With reciprocity, firms are able to counterbalance the influence of partnership opportunism and its effects on financial performance by themselves. Financial benefits achieved by opportunistically striving after economic self-interest are eliminated, in case the deprived partner attempts to opportunistically adjust its inequitable position. Our results show that opportunistic firms can benefit from immediate improvements in individual financial performance, but only until the disadvantaged partner has a counter opportunistic reaction. When the opportunism of both partners is at parity, firms cannot generate any individual increases in financial performance by acting opportunistically, highlighting the significance of bilateral opportunism for B2B business scholars. Originality This paper enhances the understanding of business relationships in industries with strong power imbalance between the buyer and the supplier firm. The authors introduce the idea of “bilateral opportunism” and “opportunistic parity” to the B2B marketing literature. These concepts shed light on the reality of business relationships, in particular in industries with high power imbalance, such as the automotive industry. Descriptive information highlights the relevance of the opportunistic parity concept.

The Effects of Data Sampling with Deep Learning and Highly Imbalanced Big Data

Training predictive models with class-imbalanced data has proven to be a difficult task. This problem is well studied, but the era of big data is producing more extreme levels of imbalance that are increasingly difficult to model. We use three data sets of varying complexity to evaluate data sampling strategies for treating high class imbalance with deep neural networks and big data. Sampling rates are varied to create training distributions with positive class sizes from 0.025%–90%. The area under the receiver operating characteristics curve is used to compare performance, and thresholding is used to maximize class performance. Random over-sampling (ROS) consistently outperforms under-sampling (RUS) and baseline methods. The majority class proves susceptible to misrepresentation when using RUS, and results suggest that each data set is uniquely sensitive to imbalance and sample size. The hybrid ROS-RUS maximizes performance and efficiency, and is our preferred method for treating high imbalance within big data problems.

ACCELERATING ECONOMIC DISTRIBUTION BY DETERMINING GROWTH POLE AREA

The Williamson index value of Central Java Province increases every year, which indicates that the development imbalance among the regions in the province belongs to the high category. Such high imbalance in a region is not positive, but indeed it is a big problem that must be solved soon by the government of Central Java Province. One appropriate effort is the acceleration of balance through adding the growth pole area. This research has several objectives among others are: 1) identifying the growth pole regions existing in Central Java Province, 2) determining the new growth pole regions in Central Java Province. This research applies the quantitative and qualitative analyses using the primary and secondary data source. The analysis uses Klassen typology, Balassa index, Scalogram, Human Development index, and overlay. The findings of this research is the regions in Central Java Province that are categorized as the old growth pole regions are Semarang City, Pekalongan City, Cilacap Regency, Karanganyar Regency, Magelang Regency, Surakarta City, and Salatiga City. While the regions that are recommended as the new growth pole regions are Semarang Regency, Tegal Regency, and Sukoharjo Regency. With the development of growth pole regions, equitable distribution of economic conditions in the Central Java region will occur and the problem of inequality will be overcome.

Author Matching Classification with Anomaly Detection Approach for Bibliomethric Repository Data

Authors name disambiguation (AND) is a complex problem in the process of identifying an author in a digital library (DL). The AND data classification process is very much determined by the grouping process and data processing techniques before entering the classifier algorithm. In general, the data pre-processing technique used is pairwise and similarity to do author matching. In a large enough data set scale, the pairwise technique used in this study is to do a combination of each attribute in the AND dataset and by defining a binary class for each author matching combination, where the unequal author is given a value of 0 and the same author is given a value of 1. The technique produces very high imbalance data where class 0 becomes 98.9% of the amount of data compared to 1.1% of class 1. The results bring up an analysis in which class 1 can be considered and processed as data anomaly of the whole data. Therefore, anomaly detection is the method chosen in this study using the Isolation Forest algorithm as its classifier. The results obtained are very satisfying in terms of accuracy which can reach 99.5%.

Author Matching Classification with Anomaly Detection Approach for Bibliomethric Repository Data

Authors name disambiguation (AND) is a complex problem in the process of identifying an author in a digital library (DL). The AND data classification process is very much determined by the grouping process and data processing techniques before entering the classifier algorithm. In general, the data pre-processing technique used is pairwise and similarity to do author matching. In a large enough data set scale, the pairwise technique used in this study is to do a combination of each attribute in the AND dataset and by defining a binary class for each author matching combination, where the unequal author is given a value of 0 and the same author is given a value of 1. The technique produces very high imbalance data where class 0 becomes 98.9% of the amount of data compared to 1.1% of class 1. The results bring up an analysis in which class 1 can be considered and processed as data anomaly of the whole data. Therefore, anomaly detection is the method chosen in this study using the Isolation Forest algorithm as its classifier. The results obtained are very satisfying in terms of accuracy which can reach 99.5%.