• Journal of Broadcast Engineering
• /
• v.21 no.4
• /
• pp.562-568
• /
• 2016

In this paper, we focus on recognition of speed-limit signs among a few types of traffic signs because speed-limit sign is closely related to safe driving of drivers. Although histogram of oriented gradient (HOG) and local binary patterns (LBP) are representative features for object recognition, these features have a weakness with respect to rotation, in that it does not consider the rotation of the target object when generating patterns. Therefore, this paper propose the fast rotation-invariant binary patterns (FRIBP) algorithm to generate a binary pattern that is robust against rotation. The proposed FRIBP algorithm deletes an unused layer of the histogram, and eliminates the shift and comparison operations in order to quickly extract the desired feature. The proposed FRIBP algorithm is successfully applied to German Traffic Sign Recognition Benchmark (GTSRB) datasets, and the results show that the recognition capabilities of the proposed method are similar to those of other methods. Moreover, its recognition speed is considerably enhanced than related works as approximately 0.47second for 12,630 test data.

• Journal of Biomedical Engineering Research
• /
• v.41 no.5
• /
• pp.185-194
• /
• 2020

Arrhythmia is basically diagnosed with the electrocardiogram (ECG) signal, however, ECG is difficult to measure and it requires expert help in analyzing the signal. On the other hand, the radial pulse can be measured with easy and uncomplicated way in daily life, and could be suitable bio-signal for the recent untact paradigm and extensible signal for diagnosis of Korean medicine based on pulse pattern. In this study, we developed an irregular radial pulse detection algorithm based on a learning model and considered its applicability as arrhythmia screening. A total of 1432 pulse waves including irregular pulse data were used in the experiment. Three data sets were prepared with minimal preprocessing to avoid the heuristic feature extraction. As classification algorithms, elastic net logistic regression, random forest, and extreme gradient boosting were applied to each data set and the irregular pulse detection performances were estimated using area under the receiver operating characteristic curve based on a 10-fold cross-validation. The extreme gradient boosting method showed the superior performance than others and found that the classification accuracy reached 99.7%. The results confirmed that the proposed algorithm could be used for arrhythmia screening. To make a fusion technology integrating western and Korean medicine, arrhythmia subtype classification from the perspective of Korean medicine will be needed for future research.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.5
• /
• pp.169-178
• /
• 2021

Recently, a bike-sharing system (BSS) has become popular as a convenient "last mile" transportation. Rebalancing of bikes is a critical issue to manage BSS because the rents and returns of bikes are not balanced by stations and periods. For efficient and effective rebalancing, accurate traffic prediction is important. Recently, cluster-based traffic prediction has been utilized to enhance the accuracy of prediction at the station-level and the clustering step is very important in this approach. In this paper, we propose a k-means based clustering algorithm that overcomes the drawbacks of the existing clustering methods for BSS; indeterministic and hardly converged. By employing the centroid initialization and using the temporal proportion of the rents and returns of stations as an input for clustering, the proposed algorithm can be deterministic and fast.

• Korean Journal of Applied Biomechanics
• /
• v.32 no.1
• /
• pp.1-8
• /
• 2022

Objective: This study proposes a methodology to analyze important variables that have a significant impact on the putting direction prediction using a machine learning-based putting direction prediction model trained with IMU sensor data. Method: Putting data were collected using an IMU sensor measuring 12 variables from 6 adult males in their 20s at K University who had no golf experience. The data was preprocessed so that it could be applied to machine learning, and a model was built using five machine learning algorithms. Finally, by comparing the performance of the built models, the model with the highest performance was selected as the proposed model, and then 12 variables of the IMU sensor were applied one by one to analyze important variables affecting the learning performance. Results: As a result of comparing the performance of five machine learning algorithms (K-NN, Naive Bayes, Decision Tree, Random Forest, and Light GBM), the prediction accuracy of the Light GBM-based prediction model was higher than that of other algorithms. Using the Light GBM algorithm, which had excellent performance, an experiment was performed to rank the importance of variables that affect the direction prediction of the model. Conclusion: Among the five machine learning algorithms, the algorithm that best predicts the putting direction was the Light GBM algorithm. When the model predicted the putting direction, the variable that had the greatest influence was the left-right inclination (Roll).

• Journal of Aerospace System Engineering
• /
• v.16 no.4
• /
• pp.109-115
• /
• 2022

The CNN (Convolution Neural Network) algorithm which combines a deep learning technique, and a computer vision technology, makes image classification feasible with the high-performance computing system. In this thesis, the CNN algorithm is applied to the classification problem, by using a typical deep learning framework of TensorFlow and machine learning techniques. The data set required for supervised learning is generated with the same type of bolts. some of which have undamaged threads, but others have damaged threads. The learning model with less quantity data showed good classification performance on detecting damage in a bolt image. Additionally, the model performance is reviewed by altering the quantity of convolution layers, or applying selectively the over and under fitting alleviation algorithm.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.463-481
• /
• 2020

Cutter suction dredgers (CSDs) are widely used in various dredging constructions such as channel excavation, wharf construction, and reef construction. During a CSD construction, the main operation is to control the swing speed of cutter to keep the slurry concentration in a proper range. However, the slurry concentration cannot be monitored in real-time, i.e., there is a "time-lag effect" in the log of slurry concentration, making it difficult for operators to make the optimal decision on controlling. Concerning this issue, a solution scheme that using real-time monitored indicators to predict current slurry concentration is proposed in this research. The characteristics of the CSD monitoring data are first studied, and a set of preprocessing methods are presented. Then we put forward the concept of "index class" to select the important indices. Finally, an ensemble learning algorithm is set up to fit the relationship between the slurry concentration and the indices of the index classes. In the experiment, log data over seven days of a practical dredging construction is collected. For comparison, the Deep Neural Network (DNN), Long Short Time Memory (LSTM), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), and the Bayesian Ridge algorithm are tried. The results show that our method has the best performance with an R² of 0.886 and a mean square error (MSE) of 5.538. This research provides an effective way for real-time predicting the slurry concentration of CSDs and can help to improve the stationarity and production efficiency of dredging construction.

• Toxicological Research
• /
• v.32 no.4
• /
• pp.289-300
• /
• 2016

Non-genotoxic carcinogens are substances that induce tumorigenesis by non-mutagenic mechanisms and long term rodent bioassays are required to identify them. Recent studies have shown that transcription profiling can be applied to develop early identifiers for long term phenotypes. In this study, we used rat liver expression profiles from the NTP (National Toxicology Program, Research Triangle Park, USA) DrugMatrix Database to construct a gene classifier that can distinguish between non-genotoxic carcinogens and other chemicals. The model was based on short term exposure assays (3 days) and the training was limited to oxidative stressors, peroxisome proliferators and hormone modulators. Validation of the predictor was performed on independent toxicogenomic data (TG-GATEs, Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System, Osaka, Japan). To build our model we performed Random Forests together with a recursive elimination algorithm (VarSelRF). Gene set enrichment analysis was employed for functional interpretation. A total of 770 microarrays comprising 96 different compounds were analyzed and a predictor of 54 genes was built. Prediction accuracy was 0.85 in the training set, 0.87 in the test set and increased with increasing concentration in the validation set: 0.6 at low dose, 0.7 at medium doses and 0.81 at high doses. Pathway analysis revealed gene prominence of cellular respiration, energy production and lipoprotein metabolism. The biggest target of toxicogenomics is accurately predict the toxicity of unknown drugs. In this analysis, we presented a classifier that can predict non-genotoxic carcinogenicity by using short term exposure assays. In this approach, dose level is critical when evaluating chemicals at early time points.

• Journal of the Korean Data Analysis Society
• /
• v.20 no.6
• /
• pp.2953-2966
• /
• 2018

Housing prices are closely related to various variables that indicate macroeconomic conditions. In this paper, empirical analysis based on data is performed referring to previous studies. Focusing on the policy interest rate among the factors affecting the housing price, the non-linear impulse responses of other variables to the interest rate shock are analyzed. Using the random forest algorithm, the variable importance scores of the macroeconomic variables presented in the previous studies are calculated. After selecting the variables through this process, the impulse responses are calculated using a model that can capture non-linearity. According to the model, the responses of housing prices to the policy rate is only significant when the rate is raised. Especially, the impulse response is amplified when the shock increases due to the non-linear characteristics that can not be captured by the traditional VAR methodology. The analysis results suggest that the interest rate as a policy instrument should be approached from a more cautious perspective.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.88-88
• /
• 2022

Detection of stress responses in crops is important to diagnose crop growth and evaluate yield. Also, the multi-spectral sensor is effectively known to evaluate stress caused by nutrient and moisture in crops or biological agents such as weeds or diseases. Therefore, in this experiment, multispectral images were taken by an unmanned aerial vehicle(UAV) under field condition. The experiment was conducted in the long-term fertilizer field in the National Institute of Crop Science, and experiment area was divided into different status of NPK(Control, N-deficiency, P-deficiency, K-deficiency, Non-fertilizer). Total 11 vegetation indices were created with RGB and NIR reflectance values using python. Variations in nutrient content in plants affect the amount of light reflected or absorbed for each wavelength band. Therefore, the objective of this experiment was to evaluate vegetation indices derived from multispectral reflectance data as input into machine learning algorithm for the classification of nutritional deficiency in rice. RandomForest model was used as a representative ensemble model, and parameters were adjusted through hyperparameter tuning such as RandomSearchCV. As a result, training accuracy was 0.95 and test accuracy was 0.80, and IPCA, NDRE, and EVI were included in the top three indices for feature importance. Also, precision, recall, and f1-score, which are indicators for evaluating the performance of the classification model, showed a distribution of 0.7-0.9 for each class.

• International Journal of Computer Science & Network Security
• /
• v.24 no.9
• /
• pp.169-185
• /
• 2024

Due to levitate and rapid growth of E-Commerce, most of the organizations are moving towards cashless transaction Unfortunately, the cashless transactions are not only used by legitimate users but also it is used by illegitimate users and which results in trouncing of billions of dollars each year worldwide. Fraud prevention and Fraud Detection are two methods used by the financial institutions to protect against these frauds. Fraud prevention systems (FPSs) are not sufficient enough to provide fully security to the E-Commerce systems. However, with the combined effect of Fraud Detection Systems (FDS) and FPS might protect the frauds. However, there still exist so many issues and challenges that degrade the performances of FDSs, such as overlapping of data, noisy data, misclassification of data, etc. This paper presents a comprehensive survey on financial fraud detection system using such data mining techniques. Over seventy research papers have been reviewed, mainly within the period 2002-2015, were analyzed in this study. The data mining approaches employed in this research includes Neural Network, Logistic Regression, Bayesian Belief Network, Support Vector Machine (SVM), Self Organizing Map(SOM), K-Nearest Neighbor(K-NN), Random Forest and Genetic Algorithm. The algorithms that have achieved high success rate in detecting credit card fraud are Logistic Regression (99.2%), SVM (99.6%) and Random Forests (99.6%). But, the most suitable approach is SOM because it has achieved perfect accuracy of 100%. But the algorithms implemented for financial statement fraud have shown a large difference in accuracy from CDA at 71.4% to a probabilistic neural network with 98.1%. In this paper, we have identified the research gap and specified the performance achieved by different algorithms based on parameters like, accuracy, sensitivity and specificity. Some of the key issues and challenges associated with the FDS have also been identified.