• 대한임베디드공학회논문지
• /
• 제16권1호
• /
• pp.17-27
• /
• 2021

In this paper, we develop a galaxy image classification method based on hand-crafted features and machine learning techniques. Additionally, we provide an empirical analysis to reveal which combination of the techniques is effective for galaxy image classification. To achieve this, we developed a framework which consists of four modules such as preprocessing, feature extraction, feature post-processing, and classification. Finally, we found that the best technique for galaxy image classification is a method to use a median filter, ORB vector features and a voting classifier based on RBF SVM, random forest and logistic regression. The final method is efficient so we believe that it is applicable to embedded environments.

• 인터넷정보학회논문지
• /
• 제23권4호
• /
• pp.45-55
• /
• 2022

암호화폐 시장의 규모는 날이 갈수록 커져가고 있으며, 대표적인 암호화폐인 비트코인의 경우 시가총액이 500조를 넘어섰다. 이에 따라 암호화폐의 가격을 예측하려는 연구도 많이 이루어졌으며, 이들은 대부분 주식가격을 예측하는 방법론과 유사성을 띄는 연구들이다. 하지만 선행연구를 비춰 봤을 때 주식가격예측과 달리 암호화폐 가격 예측은 머신러닝의 정확도가 우위에 있는 사례가 많다는 점, 개념적으로 주식과 달리 암호화폐는 소유로 인한 수동적 소득이 없다는 점, 통계적으로 시가총액 대비 하루 거래량의 비율을 살펴봤을 때 암호화폐가 주식 대비 최소 3배이상 높다는 점이 도출되었다. 이를 통해 암호화폐 가격 예측 연구에는 주식 가격 예측과 다른 방법론이 적용되어야 함을 본 논문에서 주장하였다. 우리는 기존에 주가 딥러닝 예측에 사용되던 워크 포워드 검증를 응용한 역순 워크 포워드 검증을 제안하였다. 역순 워크 포워드 검증은 워크 포워드 검증과 달리 검증 데이터셋을 테스트 데이터셋에 시계열상으로 바로 앞에 부분으로 고정시켜놓고, 훈련데이터를 훈련 데이터셋에 시계열상으로 바로 앞 부분부터 서서히 훈련 데이터셋의 크기를 늘려가면서 검증에 대한 정확도를 측정한다. 측정된 모든 검증 정확도 중 가장 높은 정확도를 보이는 훈련 데이터셋의 크기에 맞춰서 훈련 데이터를 절삭시킨 뒤 검증 데이터와 합쳐서 실험 데이터에 대한 정확도를 측정하였다. 분석모델로는 로지스틱 회귀분석과 SVM을 사용했으며, 우리가 제안한 역순 워크 포워드 검증의 신뢰성을 위해서 분석 모델 내부적으로도 L1, L2, rbf, poly등의 다양한 알고리즘과 정규화 파라미터를 적용하였다. 그 결과 모든 분석모델에서 기존 연구보다 향상된 정확도를 보임이 확인되었으며, 평균적으로도 1.23%p의 정확도 상승을 보였다. 선행연구를 통해 암호화폐 가격 예측의 정확도가 대부분 50%~60%사이에서 머무르는 걸 감안할 때 이는 상당한 정확도 개선이다.

• 한국멀티미디어학회논문지
• /
• 제19권8호
• /
• pp.1530-1537
• /
• 2016

Recently, solar power generation shows the significant growth in the renewable energy field. Using the short-term prediction, it is possible to control the electric power demand and the power generation plan of the auxiliary device. However, a short-term prediction can be used when you know the weather forecast. If it is not possible to use the weather forecast information because of disconnection of network at the island and the mountains or for security reasons, the accuracy of prediction is not good. Therefore, in this paper, we proposed a system capable of short-term prediction of solar power generation amount by using only the weather information that has been collected by oneself. We used temperature, humidity and insolation as weather information. We have applied a moving average to each information because they had a characteristic of time series. It was composed of min, max and average of each information, differences of mutual information and gradient of it. An artificial neural network, SVM and RBF Network model was used for the prediction algorithm and they were combined by Ensemble method. The results of this suggest that using a moving average during pre-processing and ensemble prediction models will maximize prediction accuracy.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2019년도 추계학술발표대회
• /
• pp.585-588
• /
• 2019

Predictions of Energy Consumption for Industries gain an important place in energy management and control system, as there are dynamic and seasonal changes in the demand and supply of energy. This paper presents and discusses the predictive models for energy consumption of the steel industry. Data used includes lagging and leading current reactive power, lagging and leading current power factor, carbon dioxide (tCO2) emission and load type. In the test set, four statistical models are trained and evaluated: (a) Linear regression (LR), (b) Support Vector Machine with radial kernel (SVM RBF), (c) Gradient Boosting Machine (GBM), (d) random forest (RF). Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) are used to measure the prediction efficiency of regression designs. When using all the predictors, the best model RF can provide RMSE value 7.33 in the test set.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권3호
• /
• pp.290-297
• /
• 2024

We investigate the efficacy of ensemble learning methods, specifically the soft voting technique, for enhancing heart disease prediction accuracy. Our study uniquely combines Logistic Regression, SVM with RBF Kernel, and Random Forest models in a soft voting ensemble to improve predictive performance. We demonstrate that this approach outperforms individual models in diagnosing heart disease. Our research contributes to the field by applying a well-curated dataset with normalization and optimization techniques, conducting a comprehensive comparative analysis of different machine learning models, and showcasing the superior performance of the soft voting ensemble in medical diagnosis. This multifaceted approach allows us to provide a thorough evaluation of the soft voting ensemble's effectiveness in the context of heart disease prediction. We evaluate our models based on accuracy, precision, recall, F1 score, and Area Under the ROC Curve (AUC). Our results indicate that the soft voting ensemble technique achieves higher accuracy and robustness in heart disease prediction compared to individual classifiers. This study advances the application of machine learning in medical diagnostics, offering a novel approach to improve heart disease prediction. Our findings have significant implications for early detection and management of heart disease, potentially contributing to better patient outcomes and more efficient healthcare resource allocation.

• Journal of Internet Technology
• /
• 제22권4호
• /
• pp.903-911
• /
• 2021

As cyber-attacks on Cyber-Physical System (CPS) become more diverse and sophisticated, it is important to quickly detect malicious behaviors occurring in CPS. Since CPS can collect sensor data in near real time throughout the process, there have been many attempts to detect anomaly behavior through normal behavior learning from the perspective of data-driven security. However, since the CPS datasets are big data and most of the data are normal data, it has always been a great challenge to analyze the data and implement the anomaly detection model. In this paper, we propose and evaluate the Clustered Deep One-Class Classification (CD-OCC) model that combines the clustering algorithm and deep learning (DL) model using only a normal dataset for anomaly detection. We use auto-encoder to reduce the dimensions of the dataset and the K-means clustering algorithm to classify the normal data into the optimal cluster size. The DL model trains to predict clusters of normal data, and we can obtain logit values as outputs. The derived logit values are datasets that can better represent normal data in terms of knowledge distillation and are used as inputs to the OCC model. As a result of the experiment, the F1 score of the proposed model shows 0.93 and 0.83 in the SWaT and HAI dataset, respectively, and shows a significant performance improvement over other recent detectors such as Com-AE and SVM-RBF.

• International Journal of Computer Science & Network Security
• /
• 제24권9호
• /
• pp.30-40
• /
• 2024

Building an accurate 3-D spatial road network model has become an active area of research now-a-days that profess to be a new paradigm in developing Smart roads and intelligent transportation system (ITS) which will help the public and private road impresario for better road mobility and eco-routing so that better road traffic, less carbon emission and road safety may be ensured. Dealing with such a large scale 3-D road network data poses challenges in getting accurate elevation information of a road network to better estimate the CO₂ emission and accurate routing for the vehicles in Internet of Vehicle (IoV) scenario. Clustering and regression techniques are found suitable in discovering the missing elevation information in 3-D spatial road network dataset for some points in the road network which is envisaged of helping the public a better eco-routing experience. Further, recently Explainable Artificial Intelligence (xAI) draws attention of the researchers to better interprete, transparent and comprehensible, thus enabling to design efficient choice based models choices depending upon users requirements. The 3-D road network dataset, comprising of spatial attributes (longitude, latitude, altitude) of North Jutland, Denmark, collected from publicly available UCI repositories is preprocessed through feature engineering and scaling to ensure optimal accuracy for clustering and regression tasks. K-Means clustering and regression using Support Vector Machine (SVM) with radial basis function (RBF) kernel are employed for 3-D road network analysis. Silhouette scores and number of clusters are chosen for measuring cluster quality whereas error metric such as MAE ( Mean Absolute Error) and RMSE (Root Mean Square Error) are considered for evaluating the regression method. To have better interpretability of the Clustering and regression models, SHAP (Shapley Additive Explanations), a powerful xAI technique is employed in this research. From extensive experiments , it is observed that SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions with an accuracy of 97.22% and strong performance metrics across all classes having MAE of 0.0346, and MSE of 0.0018. On the other hand, the ten-cluster setup, while faster in SHAP analysis, presented challenges in interpretability due to increased clustering complexity. Hence, K-Means clustering with K=4 and SVM hybrid models demonstrated superior performance and interpretability, highlighting the importance of careful cluster selection to balance model complexity and predictive accuracy.

• 지적과 국토정보
• /
• 제48권1호
• /
• pp.169-186
• /
• 2018

급속한 고령화 사회에서 노인인구가 갖는 사회적 영향력은 더욱 가속화될 것이다. 본 연구에서는 인공지능 방법론 중 머신러닝, 인공신경망, 국토정보 분석을 통해 노인복지 취약지구를 추출하는 방법론을 정립하는데 목적을 두었다. 분석방향 정립을 위해 65세 이상 노인, 공무원, 노인복지 시설물 담당자와 인터뷰 후 방향을 설정하였다. 경기도 용인시를 대상으로 500 m 공간단위 벡터 기반 격자에 15분 이내 지리적 거리 수용력, 노인복지 향유도, 공시지가, 이동통신 기반 노인활동을 지표로 설정하였다. 10단계 군집형성 후 모의학습 결과 RBF 커널 알고리즘을 활용한 머신러닝 서포트 벡터머신에서 83.2%의 예측정확도가 나타났다. 그리고 역전파 알고리즘을 활용한 인공신경망에서 높은 상관성 결과(0.63)가 나타났다. 변수간 공간적 자기상관성을 분석하기 위해 지리적 가중회귀분석을 수행했다. 분석결과 결정계수가 70.1%로 모형으로 나타나 설명력이 우수한 것으로 나타났다. 변수의 공간적 이상값 여부와 분포패턴을 검토하기 위해 국지적인 공간적 자기상관성 지수인 Moran's I 계수와 Getis-Ord Gi 계수를 분석하였다. 분석결과 용인시 신도시인 수지 기흥구에서 노인복지 취약지구가 발생하는 특성을 보였다. 본 연구의 인공지능 모의방법과 국토정보 분석의 연계는 최근 정부의 지역여건을 고려한 노인복지 불균형을 해결하는데 활용될 수 있을 것이다.

• 지능정보연구
• /
• 제22권1호
• /
• pp.187-204
• /
• 2016

소셜 미디어를 이용하는 사용자들이 직접 작성한 의견 혹은 리뷰를 이용하여 상호간의 교류 및 정보를 공유하게 되었다. 이를 통해 고객리뷰를 이용하는 오피니언마이닝, 웹마이닝 및 감성분석 등 다양한 연구분야에서의 연구가 진행되기 시작하였다. 특히, 감성분석은 어떠한 토픽(주제)를 기준으로 직접적으로 글을 작성한 사람들의 태도, 입장 및 감성을 알아내는데 목적을 두고 있다. 고객의 의견을 내포하고 있는 정보 혹은 데이터는 감성분석을 위한 핵심 데이터가 되기 때문에 토픽을 통한 고객들의 의견을 분석하는데 효율적이며, 기업에서는 소비자들의 니즈에 맞는 마케팅 혹은 투자자들의 시장동향에 따른 많은 투자가 이루어지고 있다. 본 연구에서는 중국의 온라인 시나 주식 포럼에서 사용자들이 직접 작성한 포스팅(글)을 이용하여 기존에 제시된 토픽들로부터 핫토픽을 선정하고 탐지하고자 한다. 기존에 사용된 감성 사전을 활용하여 토픽들에 대한 감성값과 극성을 분류하고, 군집분석을 통해 핫토픽을 선정하였다. 핫토픽을 선정하기 위해 k-means 알고리즘을 이용하였으며, 추가로 인공지능기법인 SOM을 적용하여 핫토픽 선정하는 절차를 제시하였다. 또한, 로짓, 의사결정나무, SVM 등의 데이터마이닝 기법을 이용하여 핫토픽 사전 탐지를 하는 감성분석을 위한 모형을 개발하여 관심지수를 통해 선정된 핫토픽과 탐지된 핫토픽을 비교하였다. 본 연구를 통해 핫토픽에 대한 정보 제공함으로써 최신 동향에 대한 흐름을 알 수 있게 되고, 주식 포럼에 대한 핫토픽은 주식 시장에서의 투자자들에게 유용한 정보를 제공하게 될 뿐만 아니라 소비자들의 니즈를 충족시킬 수 있을 것이라 기대된다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권2호
• /
• pp.483-503
• /
• 2014

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed. First, we accumulate global activities and construct motion history image (MHI) for both RGB and depth channels respectively to encode the dynamics of one action in different modalities, and then different action descriptors are extracted from depth and RGB MHI to represent global textual and structural characteristics of these actions. Specially, average value in hierarchical block, GIST and pyramid histograms of oriented gradients descriptors are employed to represent human motion. To demonstrate the superiority of the proposed method, we evaluate them by KNN, SVM with linear and RBF kernels, SRC and CRC models on DHA dataset, the well-known dataset for human action recognition. Large scale experimental results show our descriptors are robust, stable and efficient, and outperform the state-of-the-art methods. In addition, we investigate the performance of our descriptors further by combining these descriptors on DHA dataset, and observe that the performances of combined descriptors are much better than just using only sole descriptor. With multimodal features, we also propose a collaborative multi-task learning method for model learning and inference based on transfer learning theory. The main contributions lie in four aspects: 1) the proposed encoding the scheme can filter the stationary part of human body and reduce noise interference; 2) different kind of features and models are assessed, and the neighbor gradients information and pyramid layers are very helpful for representing these actions; 3) The proposed model can fuse the features from different modalities regardless of the sensor types, the ranges of the value, and the dimensions of different features; 4) The latent common knowledge among different modalities can be discovered by transfer learning to boost the performance.