检索结果-内蒙古大学图书馆

4th International Conference on Computing and Applied Informatics 2019, ICCAI 2019

作者： Hasdyna, N. Sianipar, B. Zamzami, E.M. Graduate Program of Computer Science Universitas Sumatera Utara Medan20155 Indonesia Department of Information Technology Faculty of Computer Science and Information Technology Universitas Sumatera Utara Medan20155 Indonesia

Data that has many attributes or higher dimensions will affect the performance of the K-NN classification algorithm. In this study, the Gain Ratio implemented for selecting and reducing the dataset attributes to form a new dataset for the classification process is carried out with the K-NN. The dataset used in this study are the Breast Cancer Coimbra dataset and Hepatitis C Virus dataset obtained from the UCI Machine Learning Repository. The results showed that the Breast Cancer Coimbra dataset, Gain Ratio can improve the performance of K-NN with average value 0.535596 TPR, TNR = 1, NPV = 0.608279, FNR = 1, FOR = 0.391721, Accuracy = 72.85%. In Hepatitis C Virus dataset also managed to improve the performance of K-NN with average value TPR = 0.665596, TNR = 0,876667, NPV=0,738279, FNR=0,88, FOR=0,521721, and Accuracy=86,25%. © Published under licence by IOP Publishing Ltd.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

The diversity of gamification evaluation in the software engineering education and industry: Trends, comparisons and gaps

arXiv

引用

arXiv 2021年

作者： Barbosa Monteiro, Rodrigo Henrique de Almeida Souza, Maurício Ronny Bezerra Oliveira, Sandro Ronaldo dos Santos Portela, Carlos de Cristo Lobato, Cesar Elias Graduate Program in Computer Science Federal University of Pará BelémPará Brazil Graduate Program in Computer Science Federal University of Lavras LavrasMinas Gerais Brazil Faculty of Information Systems Federal University of Pará BelémPará Brazil

Context: gamification has been used to motivate and engage participants in software engineering education and practice activities. Problem: There is a significant demand for empirical studies for the understanding of the impacts and efficacy of gamification. However, the lack of standard procedures and models for the evaluation of gamification is a challenge for the design, comparison, and report of results related to the assessment of gamification approaches and its effects. Goal: The goal of this study is to identify models and strategies for the evaluation of gamification reported in the literature. Method: To achieve this goal, we conducted a systematic mapping study to investigate strategies for the evaluation of gamification in the context of software engineering. We selected 100 primary studies on gamification in software engineering (from 2011 to 2020). We categorized the studies regarding the presence of evaluation procedures or models for the evaluation of gamification, the purpose of the evaluation, the criteria used, the type of data, instruments, and procedures for data analysis. Results: Our results show that 64 studies report procedures for the evaluation of gamification. However, only three studies actually propose evaluation models for gamification. We observed that the evaluation of gamification focuses on two aspects: the evaluation of the gamification strategy itself, related to the user experience and perceptions;and the evaluation of the outcomes and effects of gamification on its users and context. The most recurring criteria for the evaluation are "engagement", "performance", "satisfaction", and "motivation". Finally, the evaluation of gamification requires a mix of subjective and objective inputs, and qualitative and quantitative data analysis approaches. Depending of the focus of the evaluation (the strategy or the outcomes), there is a predominance of a type of data and analysis. © 2021, CC BY-NC-ND.

关键词： Software engineering

来源：评论

学校读者我要写书评

暂无评论

Customer churn prediction in telecommunication company

引用

AIP Conference Proceedings 2023年第1期2510卷

作者： Dota Biomantoro Gede Putra Kusuma Computer Science Department BINUS Graduate Program - Master of Computer Science Bina Nusantara University Jakarta Indonesia 11480

Churn prediction methods are widely used to anticipate customer churn from services provided by a company for some reasons. This study aims to develop an optimal churn prediction model based on customer data from a telecommunication company in Indonesia. The model development and evaluation processes are performed by following the Cross-Industry Standard Process for Data Mining (CRISP-DM), which consist of business understanding, data understanding, data preparation, modelling, and evaluation. Various combination of data preparation and modelling methods have been evaluated. The evaluation results show that the combination of feature selection and prediction model yields better results compared to prediction model without feature selection. The highest accuracy is achieved by Random Forrest at 97.82%, which is followed by Decision Tree at 97.06%, and Naive Bayes at 90.62%. This result indicates that a prediction model can be reliably used to predict customer churn in a telecommunication company.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Impact of Image Enhancement for Osteoporosis Detection Based on Deep Learning Algorithm

Impact of Image Enhancement for Osteoporosis Detection Based...

引用

computer System, Information Technology, and Electrical Engineering (COSITE), International Conference on

作者： Riza Dianil Iman Roslidar Roslidar Kahlil Muchtar Amalia Amalia Mohammad Syaryadhi Alfatirta Mufti Novi Maulina Department of Electrical and Computer Engineering Universitas Syiah Kuala Banda Aceh Indonesia Department of Computer Science Universitas Sumatera Utara Medan Indonesia Graduate School of System Designs Tokyo Metropolitan University Tokyo Japan Doctorate Student of Doctoral Program in Medical Sciences Faculty of Medicine Universitas Syiah Kuala Banda Aceh Indonesia

Osteoporosis can be defined as a degenerative disease with reduced bone mass and changes in bone architecture that can lead to bone fragility and the risk of fractures. This abnormality can be indicated by the bone density which in visual can be determined using X-Ray images. However, X-Ray images are susceptible to noise, while in image analysis image contrast affects deep learning abilities. Hence, the CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm is used as a contrast enhancement technique in X-Ray images. This study aims to build a deep learning model using the CLAHE-enhanced image dataset with the ResNet-50 and ResNet-101 architectures. The model was built using two different datasets, namely the original image dataset and the CLAHE-enhanced image dataset. The result shows that the highest performance is given by the ResNet-101 model using the CLAHE-enhanced image dataset with an accuracy rate of 96%, precision of 95%, specificity of 95%, recall of 97% and an Fl-score of 96%, respectively. By using the CLAHE algorithm, the resulting image has high contrast and looks better at displaying features in the image so as to produce better model performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Calibration of BLE beacons and its impact on distance estimation using the log-distance path loss model

Calibration of BLE beacons and its impact on distance estima...

引用

Latin-American Symposium on Dependable Computing, LADC

作者： Leonardo Vanzin Marcio Seiji Oyamada Computer Science Course UNIOESTE Western State University of Parana Cascavel Brazil Graduate Program in Computer Science UNIOESTE Western State University of Parana Cascavel Brazil

ISBN: (纸本)9781665478328

The indoor positioning system is used to determine the location of people or objects in an enclosed space. It can be used in many applications such as navigation or even active marketing to provide a better experience to customers. There are various solutions to perform indoor localization. In this paper, the focus is on the use of Bluetooth Low Energy beacons to determine the position of the customer or object in the room. Therefore, the goal of this work is to use low-cost HM-10 BLE boards to test their accuracy and use as beacons. After collecting and analyzing the data, it was determined that HM-10 is a viable solution for beacons. When using the log-distance path loss model to determine distance, the experiments show variations between different HM-10 boards, suggesting the need for calibration to obtain more accurate results.

关键词： Location awareness Power demand Fluctuations Costs Bluetooth Navigation Estimation

来源：评论

学校读者我要写书评

暂无评论

Improving Candle Direction Classification in Forex Market Using Support Vector Machine with Hyperparameters Tuning

Improving Candle Direction Classification in Forex Market Us...

引用

International Conference on Informatics and Computing (ICIC)

作者： Raymond Sunardi Oetama Yaya Heryadi Lukas Wayan Suparta Doctor of Computer Science Bina Nusantara University Jakarta Indonesia Computer Science Department Binus Graduate Program Doctor of Computer Science Bina Nusantara University Jakarta Indonesia Cognitive Engineering Research Group Universitas Katolik Indonesia Atma Jaya Jakarta Indonesia Department of Electrical Engineering Faculty of Industrial Technology Institut Teknologi Nasional Yogyakarta Indonesia

ISBN: (纸本)9798350345728

Foreign Exchange market is the world's largest daily currency turnover. Two of the popular currencies Euro and Pound sterling traded against the US Dollar. Since the Russia and Ukraine war started in February 2022, their exchange rates decrease to the lowest rate ever. Even though the general trend is bearish, several daily candles increase for some days making challenges for forex analysts. To solve this problem, classification is applied. The data is labeled downward and upward. By utilizing Linear Kernel and Radial Basis Function (RBF) Kernel-based Support Vector Machines (SVM), the candle direction can be classified and optimized by tuning the Hyperparameters. The accuracy of candle direction classifications are highly improved. After tuning, in general, classification using Linear Models can outperform RBF Models. The best accuracy found on the Pound sterling against US Dollar by using the Linear model is 98.11% and the accuracy becomes 100% on data testing at a ratio of 70:30. Whilst for the Euro against the US Dollar, the best accuracy found the same for both Linear and RBF models on a ratio of 80:20 at 97.53%. However, on data testing, it decreases to 94.51% for Linear Model and 93.41% using RBF Model. The implication of this study is SVM can successfully classify candle direction on pairs in the Forex Market that are affected by a big event that comes for such a long period as long as the hyperparameter is tuned.

关键词： Support vector machines Training Exchange rates Market research Data models Kernel Informatics

来源：评论

学校读者我要写书评

暂无评论

Out-of-Candidate Rectification for Weakly Supervised Semantic Segmentation

Out-of-Candidate Rectification for Weakly Supervised Semanti...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Zesen Cheng Pengchong Qiao Kehan Li Siheng Li Pengxu Wei Xiangyang Ji Li Yuan Chang Liu Jie Chen School of Electronic and Computer Engineering Peking University Shenzhen China AI for Science (AI4S)-Preferred Program Peking University Shenzhen Graduate School China Peng Cheng Laboratory Shenzhen China Tsinghua University Sun Yat-Sen University

Weakly supervised semantic segmentation is typically inspired by class activation maps, which serve as pseudo masks with class-discriminative regions highlighted. Although tremendous efforts have been made to recall precise and complete locations for each class, existing methods still commonly suffer from the unsolicited Out-of-Candidate (OC) error predictions that do not belong to the label candidates, which could be avoidable since the contradiction with image-level class tags is easy to be detected. In this paper, we develop a group ranking-based Out-of-f;Candidate Rectification (OCR) mechanism in a plug-and-play fashion. Firstly, we adaptively split the semantic categories into In-Candidate (IC) and OC groups for each OC pixel according to their prior annotation correlation and posterior prediction correlation. Then, we derive a differentiable rectification loss to force OC pixels to shift to the IC group. Incorporating OCR with seminal baselines (e.g., AffinityNet, SEAM, MCTformer), we can achieve remarkable performance gains on both Pascal VOC (+3.2%, +3.3%, +0.8% mIoU) and MS COCO (+1.0%, +1.3%, +0.5% mIoU) datasets with negligible extra training overhead, which jus-tifies the effectiveness and generality of OCR. † † Ŋ ***/sennnnn/Out-of-Candidate-Rectification

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep peak property learning for efficient chiral molecules ECD spectra prediction

arXiv

引用

arXiv 2024年

作者： Li, Hao Long, Da Yuan, Li Tian, Yonghong Wang, Xinchang Mo, Fanyang School of Electronic and Computer Engineering Peking University Shenzhen China -Preferred Program Peking University Shenzhen Graduate School Shenzhen China College of Chemistry and Chemical Engineering Xiamen University Xiamen361005 China Peng Cheng Laboratory Shenzhen China School of Materials Science and Engineering Peking University Beijing China

Chiral molecule assignation is crucial for asymmetric catalysis, functional materials, and the drug industry. The conventional approach requires theoretical calculations of electronic circular dichroism (ECD) spectra, which is time-consuming and costly. To speed up this process, we have incorporated deep learning techniques for the ECD prediction. We first set up a large-scale dataset of Chiral Molecular ECD spectra (CMCDS) with calculated ECD spectra. We further develop the ECDFormer model, a Transformer-based model to learn the chiral molecular representations and predict corresponding ECD spectra with improved efficiency and accuracy. Unlike other models for spectrum prediction, our ECDFormer creatively focused on peak properties rather than the whole spectrum sequence for prediction, inspired by the scenario of chiral molecule assignation. Specifically, ECDFormer predicts the peak properties, including number, position, and symbol, then renders the ECD spectra from these peak properties, which significantly outperforms other models in ECD prediction, Our ECDFormer reduces the time of acquiring ECD spectra from 1-100 hours per molecule to 1.5s. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Low-Energy and Reduced-Area Hardware Architecture for the Versatile Video Coding FME

Low-Energy and Reduced-Area Hardware Architecture for the Ve...

引用

Symposium on Integrated Circuits and Systems Design (SBCCI)

作者： Vanio Rodrigues Filho Ismael Seidel Nicole Citadin Marcio Monteiro Mateus Grellert José Luis Güntzel Dept. of Informatics and Statistics (INE) Embedded Computing Lab. (ECL) Computer Science Graduate Program (PPGCC) Federal University of Santa Catarina (UFSC) Florianópolis Brazil Federal University of Rio Grande do Sul (UFRGS) Porto Alegre Brazil

The Versatile Video Coding (VVC) standard incorporates a number of new tools to significantly improve the coding efficiency over its predecessor, the High Efficiency Video Coding (HEVC), at the cost of increased complexity. Hence, to meet real-time and energy efficiency requirements, portable devices must adopt techniques to effectively reduce the VVC complexity, including customized hardware accelerators. This work presents a low-energy hardware architecture design for the Fractional Motion Estimation (FME), which is amongst the most critical VVC and HEVC encoder steps. The proposed architecture reduces the FME complexity by operating over a subset of candidates selected to reduce the energy and area demands without compromising too much the coding efficiency. At the average BD-Rate of only 0.34% and 0.28% for the Low Delay with P slices only (LD-P) and Random Access (RA) configurations, respectively, the proposed architecture achieves a 75.5% energy consumption reduction when compared to a baseline FME architecture while occupying only 36.6 % of the area.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Hybrid descriptor definition for content based image classification using fusion of handcrafted features to convolutional neural network features

引用

International Journal of Information Technology (Singapore) 2021年第4期13卷 1365-1374页

作者： Das, Rik Kumari, Khusbu De, Sourav Manjhi, P.K. Thepade, Sudeep Department of Post Graduate Program in Information Technology Xavier Institute of Social Service Ranchi Jharkhand India Department of Computer Applications Yogoda Satsanga Mahavidyalaya Ranchi Jharkhand India Department of Computer Science & Engineering Cooch Behar Government Engineering College Cooch Behar India Department of Mathematics Vinoba Bhave University Hazaribag Jharkhand India Department of Computer Engineering Pimpri Chinchwad College of Engineering Pune India

The efficacy of content-based image classification is dependent on the richness of the feature vectors extracted from the image data. Traditional feature extraction techniques highlight single low level image characteristics like shape, size, texture, color etc. for feature generation. The process often fails to extract meaningful descriptors since considering a single image characteristic will ignore other rich properties of image contents. Mass adoption of Convolutional Neural Network (CNN) has significantly improved classification performances for content-based image data. Recent literature has documented high level of precision for image classification by carrying out transfer learning with open source pre trained Convolutional Neural Network (CNN). However, the concept of transfer learning experiences major setback in case of limited training data due to overfitting of training instances during finetuning. This work has identified this challenge and has attempted to capture probability distribution of input images to a pre trained Convolution Neural Network by utilizing it as a fixed weight feature extractor. The proposed approach has addressed the overfitting problem by removing the finetuning step involved in transfer learning. Further investigation on robust descriptor definition is carried out by concatenating the pre-trained CNN features to handcrafted features. The hybrid architecture has encouraging outcomes that have outclassed the classification accuracies of state-of-the-art handcrafted techniques. © 2021, Bharati Vidyapeeth's Institute of computer Applications and Management.

关键词： Augmented feature learning Feature fusion Handcrafted features Image classification Pre trained Convolutional Neural Network (CNN) Representation learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：