• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.12
• /
• pp.445-452
• /
• 2022

Class distribution of unbalanced data is an important part of the digital world and is a significant part of cybersecurity. Abnormal activity of unbalanced data should be found and problems solved. Although a system capable of tracking patterns in all transactions is needed, machine learning with disproportionate data, which typically has abnormal patterns, can ignore and degrade performance for minority layers, and predictive models can be inaccurately biased. In this paper, we predict target variables and improve accuracy by combining estimates using Synthetic Minority Oversampling Technique (SMOTE) and Light GBM algorithms as an approach to address unbalanced datasets. Experimental results were compared with logistic regression, decision tree, KNN, Random Forest, and XGBoost algorithms. The performance was similar in accuracy and reproduction rate, but in precision, two algorithms performed at Random Forest 80.76% and Light GBM 97.16%, and in F1-score, Random Forest 84.67% and Light GBM 91.96%. As a result of this experiment, it was confirmed that Light GBM's performance was similar without deviation or improved by up to 16% compared to five algorithms.

• Journal of Broadcast Engineering
• /
• v.28 no.2
• /
• pp.238-241
• /
• 2023

Recently deep learning models are widely used for various tasks such as facial recognition and face editing. Their training process often involves a dataset with imbalanced age distribution. It is because some age groups (teenagers and middle age) are more socially active and tends to have more data compared to the less socially active age groups (children and elderly). This imbalanced age distribution may negatively impact the deep learning training process or the model performance when tested against those age groups with less data. To this end, we propose an age-controllable face synthesis technique using a feature decomposition to classify age from facial images which can be utilized to synthesize novel data to balance out the age distribution. We perform extensive qualitative and quantitative evaluation on our proposed technique using the FFHQ dataset and we show that our method has better performance than existing method.

• The KIPS Transactions:PartD
• /
• v.19D no.1
• /
• pp.21-28
• /
• 2012

Predication of protein interaction sites for monomer structures can reduce the search space for protein docking and has been regarded as very significant for predicting unknown functions of proteins from their interacting proteins whose functions are known. In the other hand, the prediction of interaction sites has been limited in crystallizing weakly interacting complexes which are transient and do not form the complexes stable enough for obtaining experimental structures by crystallization or even NMR for the most important protein-protein interactions. This work reports the calculation of 3D surface patches of complex structures and their properties and a machine learning approach to build a predictive model for the 3D surface patches in interaction and non-interaction sites using support vector machine. To overcome classification problems for class imbalanced data, we employed an under-sampling technique. 9 properties of the patches were calculated from amino acid compositions and secondary structure elements. With 10 fold cross validation, the predictive model built from SVM achieved an accuracy of 92.7% for classification of 3D patches in interaction and non-interaction sites from 147 complexes.

• Journal of the Korean Society of Safety
• /
• v.36 no.3
• /
• pp.31-39
• /
• 2021

In Korea, more than half of work-related fatalities have occurred on construction sites. To reduce such occupational accidents, safety inspection by government agencies is essential in construction sites that present a high risk of serious accidents. To address this issue, this study developed risk prediction models of serious accidents in construction sites using five machine learning methods: support vector machine, random forest, XGBoost, LightGBM, and AutoML. To this end, 15 proactive information (e.g., number of stories and period of construction) that are usually available prior to construction were considered and two over-sampling techniques (SMOTE and ADASYN) were used to address the problem of class-imbalanced data. The results showed that all machine learning methods achieved 0.876~0.941 in the F1-score with the adoption of over-sampling techniques. LightGBM with ADASYN yielded the best prediction performance in both the F1-score (0.941) and the area under the ROC curve (0.941). The prediction models revealed four major features: number of stories, period of construction, excavation depth, and height. The prediction models developed in this study can be useful both for government agencies in prioritizing construction sites for safety inspection and for construction companies in establishing pre-construction preventive measures.

• Information Systems Review
• /
• v.24 no.4
• /
• pp.97-118
• /
• 2022

As generative adversarial network (GAN) based oversampling techniques have achieved impressive results in class imbalance of unstructured dataset such as image, many studies have begun to apply it to solving the problem of imbalance in structured dataset. However, these studies have failed to reflect the characteristics of structured data due to changing the data structure into an unstructured data format. In order to overcome the limitation, this study adapted CycleGAN to reflect the characteristics of structured data, and proposed hybridization of synthetic minority oversampling technique (SMOTE) and the adapted CycleGAN. In particular, this study tried to overcome the limitations of existing studies by using a one-dimensional convolutional neural network unlike previous studies that used two-dimensional convolutional neural network. Oversampling based on the method proposed have been experimented using various datasets and compared the performance of the method with existing oversampling methods such as SMOTE and adaptive synthetic sampling (ADASYN). The results indicated the proposed hybrid oversampling method showed superior performance compared to the existing methods when data have more dimensions or higher degree of imbalance. This study implied that the classification performance of oversampling structured data can be improved using the proposed hybrid oversampling method that considers the characteristic of structured data.

• Journal of Electrical Engineering and Technology
• /
• v.9 no.6
• /
• pp.2098-2106
• /
• 2014

In the area of clustering, there are numerous approaches to construct clusters in the input space. For regression problem, when forming clusters being a part of the overall model, the relationships between the input space and the output space are essential and have to be taken into consideration. Conditional Fuzzy C-Means (c-FCM) clustering offers an opportunity to analyze the structure in the input space with the mechanism of supervision implied by the distribution of data present in the output space. However, like other clustering methods, c-FCM focuses on the distribution of the data. In this paper, we introduce a new method, which by making use of the ambiguity index focuses on the boundaries of the clusters whose determination is essential to the quality of the ensuing classification procedures. The introduced design is illustrated with the aid of numeric examples that provide a detailed insight into the performance of the fuzzy classifiers and quantify several essentials design aspects.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.327-338
• /
• 2022

Classification models for human action recognition require robust features and large training sets for good generalization. However, data augmentation methods are employed for imbalanced training sets to achieve higher accuracy. These samples generated using data augmentation only reflect existing samples within the training set, their feature representations are less diverse and hence, contribute to less precise classification. This paper presents new data augmentation and action representation approaches to grow training sets. The proposed approach is based on two fundamental concepts: virtual video generation for augmentation and representation of the action videos through robust features. Virtual videos are generated from the motion history templates of action videos, which are convolved using a convolutional neural network, to generate deep features. Furthermore, by observing an objective function of the genetic algorithm, the spatiotemporal features of different samples are combined, to generate the representations of the virtual videos and then classified through an extreme learning machine classifier on MuHAVi-Uncut, iXMAS, and IAVID-1 datasets.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.808-813
• /
• 2022

This study presents the problems of data imbalance, varying difficulties across target objects, and small objects in construction object segmentation for far-field monitoring and utilize compound loss functions to address it. Construction site scenes of assembling scaffolds were analyzed to test the effectiveness of compound loss functions for five construction object classes---workers, hardhats, harnesses, straps, hooks. The challenging problem was mitigated by employing a focal and Jaccard loss terms in the original loss function of LinkNet segmentation model. The findings indicates the importance of the loss function design for model performance on construction site scenes for far-field monitoring.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.5
• /
• pp.1886-1908
• /
• 2020

With the technical advances, the amount of big data is increasing day-by-day such that the traditional software tools face a burden in handling them. Additionally, the presence of the imbalance data in big data is a massive concern to the research industry. In order to assure the effective management of big data and to deal with the imbalanced data, this paper proposes a new indexing algorithm for retrieving big data in the MapReduce framework. In mappers, the data clustering is done based on the Sparse Fuzzy-c-means (Sparse FCM) algorithm. The reducer combines the clusters generated by the mapper and again performs data clustering with the Sparse FCM algorithm. The two-level query matching is performed for determining the requested data. The first level query matching is performed for determining the cluster, and the second level query matching is done for accessing the requested data. The ranking of data is performed using the proposed Monarch chaotic whale optimization algorithm (M-CWOA), which is designed by combining Monarch butterfly optimization (MBO) [22] and chaotic whale optimization algorithm (CWOA) [21]. Here, the Parametric Enabled-Similarity Measure (PESM) is adapted for matching the similarities between two datasets. The proposed M-CWOA outperformed other methods with maximal precision of 0.9237, recall of 0.9371, F1-score of 0.9223, respectively.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.225-234
• /
• 2019

Deep learning has become increasingly popular in both academic and industrial areas nowadays. Various domains including pattern recognition, Computer vision have witnessed the great power of deep neural networks. However, current studies on deep learning mainly focus on quality data sets with balanced class labels, while training on bad and imbalanced data set have been providing great challenges for classification tasks. We propose in this paper a method of data analysis-based data reduction techniques for selecting good and diversity data samples from a large dataset for a deep learning model. Furthermore, data sampling techniques could be applied to decrease the large size of raw data by retrieving its useful knowledge as representatives. Therefore, instead of dealing with large size of raw data, we can use some data reduction techniques to sample data without losing important information. We group PCB characters in classes and train deep learning on the ResNet56 v2 and SENet model in order to improve the classification performance of optical character recognition (OCR) character classifier.