检索结果-内蒙古大学图书馆

European Signal processing conference (EUSIPCO)

作者： Hamed Yazdanpanah Ana Clara Teixeira Vaishali Marar Aline Pezente Mohammad M. Ghassemi Traive Inc. Brookline MA USA

ISBN: (数字)9789464593617

ISBN: (纸本)9798331519773

Reliable forecasting of regional commodity prices is crucial for agribusiness risk management. Our paper introduces data-Selective Normalized Least-Mean-Square (DS-NLMS), a new adaptive learning algorithm applied to 90-day ahead daily soybean price forecasts. Trained until March 2020, the model was tested on unseen data to evaluate its robustness against COVID-19's impact on prices in 35 Brazil-ian cities. The DS-NLMS outperformed benchmarks like Long Short-Term Memory network (LSTM) and the Chicago Board of Trade futures (CBOT), achieving a Mean Absolute Percentage Error (MAPE) of 12.90%, significantly lower than LSTM's 25.51% and CBOT's 13.74%. While the performance of DS-NLMS in comparison with CBOT is less definitive at a 90-day horizon, this comparison shows statistical significance at a 30-day horizon, with a $p$ -value of < 1.24 × 10– 7 . The DS-NLMS is also the most computationally efficient model, making it highly practical for real-world applications.

关键词： Computational modeling Urban areas Signal processing algorithms Europe Signal processing Robustness data models Risk management Forecasting Long short term memory

来源：评论

学校读者我要写书评

暂无评论

RTAD-TP: Real-Time Anomaly Detection Algorithm for Univariate Time Series data Based on Two-Parameter Estimation 1

引用

16th international conference on Knowledge Science, Engineering and Management (KSEM)

作者： Fan, Qiyun Tang, Yan Ding, Xiaoming Huangfu, Qianglong Ding, Peihao Southwest Univ Sch Comp & Informat Sci Chongqing Peoples R China

ISBN: (数字)9783031402838

ISBN: (纸本)9783031402821;9783031402838

The anomaly detection of univariate time series data is becoming increasingly important, as early detection of anomalies is valuable in practical applications. However, due to the continuous influx of data streams and the dynamic changes in data patterns, real-time anomaly detection still poses challenges. algorithms such as SPOT, DSPOT, and FluxEV are efficient unsupervised anomaly detection algorithms for data streams, but their detection performance still needs to be improved when processing large-scale data streams. To address this, we propose the Real-Time Anomaly Detection Algorithm for Univariate Time Series data Based on Two-Parameter Estimation (RTAD-TP), a fast and effective unsupervised real-time anomaly detection algorithm. We calculate the residual between the current value and the predicted value using Exponential Moving Average (EMA), and apply extreme value distribution to determine the threshold for the residual. In addition, we use two-parameter estimation to improve the speed and accuracy of parameter estimation in automatic thresholding, addressing the limitations of SPOT, DSPOT, and FluxEV. Experimental results show that the RTAD-TP algorithm has better detection performance than the baseline algorithm.

关键词： Univariate time series data real-time anomaly detection unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Multi-domain Recommendation with Embedding Disentangling and Domain Alignment 23

Multi-domain Recommendation with Embedding Disentangling and...

引用

32nd ACM international conference on Information and Knowledge Management (CIKM)

作者： Ning, Wentao Yan, Xiao Liu, Weiwen Cheng, Reynold Zhang, Rui Tang, Bo Univ Hong Kong Hong Kong Peoples R China Southern Univ Sci & Technol Hong Kong Peoples R China Huawei Noahs Ark Lab Montreal PQ Canada Univ Hong Kong Dept Comp Sci Hong Kong Peoples R China Univ Hong Kong Musketeers Fdn Inst Data Sci Hong Kong Peoples R China Guangdong Hong Kong Macau Joint Lab Hong Kong Peoples R China Southern Univ Sci & Technol Res Inst Trustworthy Autonomous Syst Shenzhen Peoples R China Southern Univ Sci & Technol Dept Comp Sci & Engn Shenzhen Peoples R China

ISBN: (纸本)9798400701245

Multi-domain recommendation (MDR) aims to provide recommendations for different domains (e.g., types of products) with overlapping users/items and is common for platforms such as Amazon, Facebook, and LinkedIn that host multiple services. Existing MDR models face two challenges: First, it is difficult to disentangle knowledge that generalizes across domains (e.g., a user likes cheap items) and knowledge specific to a single domain (e.g., a user likes blue clothing but not blue cars). Second, they have limited ability to transfer knowledge across domains with small overlaps. We propose a new MDR method named EDDA with two key components, i.e., embedding disentangling recommender and domain alignment, to tackle the two challenges respectively. In particular, the embedding disentangling recommender separates both the model and embedding for the inter-domain part and the intra-domain part, while most existing MDR methods only focus on model-level disentangling. The domain alignment leverages random walks from graph processing to identify similar user/item pairs from different domains and encourages similar user/item pairs to have similar embeddings, enhancing knowledge transfer. We compare EDDA with 12 state-of-the-art baselines on 3 real datasets. The results show that EDDA consistently outperforms the baselines on all datasets and domains. All datasets and codes are available at https://***/Stevenn9981/EDDA.

关键词： multi-domain recommendation disentanglement alignment

来源：评论

学校读者我要写书评

暂无评论

data ANALYTICS IN FARMING: RICE PRICE PREDICTION IN ANDHRA PRADESH 5

DATA ANALYTICS IN FARMING: RICE PRICE PREDICTION IN ANDHRA P...

引用

5th international conference on Multimedia, Signal processing and Communication Technologies (IMPACT)

作者： Reddy, Pundru Chandra Shaker Suryanarayana, G. Prakash, L. N. C. K. Yadala, Sucharitha REVA Univ Sch Comp & Informat Technol Bengalure India Vardhaman Coll Engn Dept Comp Sci & Engn Hyderabad India CVR Coll Engn Dept Comp Sci & Engn Hyderabad India CMR Inst Technol Dept Comp Sci & Engn Hyderabad India

ISBN: (纸本)9781665476478

The variances regarding prices of horticultural commodities adversely affect the nations GDP. The ranchers are genuinely and monetarily influenced as their long stretches of difficult work go to no end. Forecast of the prices might help the farming production network in settling on vital choices in limiting and dealing with the danger in value variances. By the way of outcome of the diminution of crop yield making due to uneven atmosphere circumstances and global warming. Rice is a significant yield developed in India, with Andhra Pradesh (AP) state being second as far as creation. At present ranchers in AP are moving from rice development to different harvests on account of worth variances and environmental variation. In this research, the costs of rice in AP remain anticipated utilizing time-series using data analytics like Machine Learning (ML) models. The prototypes SARIMA, Holt-Winter's Periodic technique, and Long Short-Term Memory (LSTM) Neural Networks (NN)be situated and utilized, and their recital was assessed dependent on the RMSE esteem on the rice database whose costs from 2001 to 2020. The conclusions demonstrate that the recommended process, LSTM neural network method is actual expectant and finest method which is suitable for the data.

关键词： Rice price prediction machine learning SARIMA LSTM

来源：评论

学校读者我要写书评

暂无评论

Medical data Clustering Based on Multi-objective Clustering Algorithm 4th

Medical Data Clustering Based on Multi-objective Clustering ...

引用

4th international conference on Machine Learning for Cyber Security, ML4CS 2022

作者： Chen, Shilian Tan, Yingsi Guo, Junkai He, Yuqin Geng, Shuang College of Management Shenzhen University Shenzhen518060 China

ISBN: (纸本)9783031201011

With the development of massive medical data, clustering algorithm becomes an effective way for medical data processing and data mining. On the one hand, it helps medical learners find effective information patterns from massive data;on the other hand, it promotes the development of medical technology and increase productivity. For traditional clustering algorithm, a single clustering index is difficult to meet people's needs of diversity and comprehensiveness. In contrast, multi-objective clustering (MOC) considers multiple objectives at the same time, and comprehensively deals with various clustering problems and standards, such as compactness, diversity of feature selection and high data dimension. Artificial bee colony algorithm (ABC) has a faster speed and embodies the idea of swarm intelligence. It imitates the optimization process of bees, and finally obtains the global optimal value. On this basis, this paper proposed a multi-objective artificial bee colony clustering algorithm (MOC-NABC) that is combined with current better-performed clustering algorithm. It takes normalized mutual information (NMI), Calinski-Harabasz (CH), Fowlkes-Mallows index (FMI) and silhouette coefficient (SC) of clustering as the final evaluation indexes. The experiment on UCI mouse protein gene dataset shows that the overall performance effect is greatly improved, e.g. compact clustering and the effective utilization of data features. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Quantum Fuzzy Least Squares Algorithm for Uncertain Environment 8th

Quantum Fuzzy Least Squares Algorithm for Uncertain Environm...

引用

8th international conference on Artificial Intelligence and Security , ICAIS 2022

作者： Yu, Shipeng Chang, Yan Zhang, Shibin Wang, Qirun School of Cybersecurity Chengdu University of Information Technology Chengdu610225 China Advanced Cryptography and System Security Key Laboratory of Sichuan Province Chengdu610255 China School of Engineering and Technology University of Hertfordshire Hertford United Kingdom

ISBN: (纸本)9783031067662

With the development of information technology, modern society has entered the era of big data. Big data means a huge amount of data, as well as the complexity and uncertainty of data. In order to find valuable data efficiently and accurately from the big data environment, researchers use machine learning algorithms to establish relevant data models to predict and classify new data. In the process of establishing the linear system model, the commonly used method is the least square method, which can easily obtain the optimal parameters of the linear system and make the established linear model more accurate. In order to deal with the fuzziness of the data itself and reduce the influence of abnormal data on the linear model, researchers put forward the fuzzy linear model. When the amount of data is increasing, the time required for traditional computers to process data and establish models is increasing. In order to reduce the time of establishing the model, a quantum fuzzy least square method is proposed on the basis of predecessors. The algorithm uses the advantages of quantum computing in data processing to improve the efficiency of fuzzy linear model establishment and reduce the model establishment time. Relevant research and analysis prove the efficiency of the algorithm. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Big data

来源：评论

学校读者我要写书评

暂无评论

Show Me the Whole World: Towards Entire Item Space Exploration for Interactive Personalized Recommendations 22

Show Me the Whole World: Towards Entire Item Space Explorati...

引用

15th ACM international conference on Web Search and data Mining (WSDM)

作者： Song, Yu Sun, Shuai Lian, Jianxun Huang, Hong Li, Yu Jin, Hai Xie, Xing Natl Engn Res Ctr Big Data Technol & Syst Wuhan Peoples R China Serv Comp Technol & Syst Lab Wuhan Peoples R China Luazhong Univ Sci & Technol Wuhan Peoples R China Microsoft Res Asia Beijing Peoples R China Meituan Grp Beijing Peoples R China

ISBN: (纸本)9781450391320

User interest exploration is an important and challenging topic in recommender systems, which alleviates the closed-loop effects between recommendation models and user-item interactions. Contextual bandit (CB) algorithms strive to make a good trade-off between exploration and exploitation so that users' potential interests have chances to expose. However, classical CB algorithms can only be applied to a small, sampled item set (usually hundreds), which forces the typical applications in recommender systems limited to candidate post-ranking, homepage top item ranking, ad creative selection, or online model selection (A/B test). In this paper, we introduce two simple but effective hierarchical CB algorithms to make a classical CB model (such as LinUCB and Thompson Sampling) capable to explore users' interest in the entire item space without limiting to a small item set. We first construct a hierarchy item tree via a bottom-up clustering algorithm to organize items in a coarse-to-fine manner. Then we propose a hierarchical CB (HCB) algorithm to explore users' interest on the hierarchy tree. HCB takes the exploration problem as a series of decision-making processes, where the goal is to find a path from the root to a leaf node, and the feedback will be back-propagated to all the nodes in the path. We further propose a progressive hierarchical CB (pHCB) algorithm, which progressively extends visible nodes which reach a confidence level for exploration, to avoid misleading actions on upper-level nodes in the sequential decision-making process. Extensive experiments on two public recommendation datasets demonstrate the effectiveness and flexibility of our methods.

关键词： Recommender System Contextual Bandit Interest Exploration

来源：评论

学校读者我要写书评

暂无评论

The Role of User Behavior Patterns in Enhancing Fraud Detection in Online Banking: A Bibliometric Analysis

The Role of User Behavior Patterns in Enhancing Fraud Detect...

引用

Technology Innovation and Its Applications (ICTIIA), international conference on

作者： Pita Silvia Qurotul Aini Efa Ayu Nabila Henderi Hardjanto Nusantoro Department of Informatic Engineering University of Raharja Tangerang Indonesia Department of Digital Business University of Raharja Tangerang Indonesia Department of Retail Management University of Raharja Tangerang Indonesia

ISBN: (数字)9798350351613

ISBN: (纸本)9798350351620

The complexity of fraud in online banking has outpaced traditional rule-based detection systems. This paper examines how user behavior patterns enhance fraud detection by integrating behavioral analytics with machine learning models to improve accuracy and reduce false positives. A bibliometric analysis using VOSviewer was conducted on 200 Google Scholar papers (2020–2024) to identify key trends. Terms like “user behavior” and “machine learning algorithms” emerged as central. Machine learning models, including neural networks and deep learning, were evaluated, with Convolutional Neural Networks (CNNs) achieving up to 95% accuracy. Challenges such as computational complexity, data privacy, and overfitting persist. Future research should explore hybrid models that integrate multiple data sources to further improve fraud detection systems. These findings align with IEEE's focus on cybersecurity and financial technology advancements.

关键词： Analytical models Accuracy Online banking Computational modeling Soft sensors Biological system modeling Neural networks data models Fraud Overfitting

来源：评论

学校读者我要写书评

暂无评论

A Statistical-Based Approach to Load Model Parameter Identification

引用

IEEE ACCESS 2021年 9卷 66915-66928页

作者： Gulakhmadov, Aminjon Tavlintsev, Alexander Pankratov, Aleksey Suvorov, Anton Kovaleva, Anastasia Lipnitskiy, Ilya Safaraliev, Murodbek Semenenko, Sergey Gubin, Pavel Dmitriev, Stepan Rasulzoda, Khusrav Chinese Acad Sci Xinjiang Inst Ecol & Geog Res Ctr Ecol & Environm Cent Asia Urumqi 830011 Peoples R China Chinese Acad Sci Xinjiang Inst Ecol & Geog State Key Lab Desert & Oasis Ecol Urumqi 830011 Peoples R China Minist Energy & Water Resources Tajikistan Dushanbe 734064 Tajikistan Ural Fed Univ Dept Automated Elect Syst Ekaterinburg 620002 Russia United Power Syst Tomsk 634041 Russia HP Inc Portland OR 97211 USA Andritz Hydro GmbH D-88212 Ravensburg Germany

In the last few years, a great number of methods for identifying the load model parameters have been proposed. This article discusses the use of statistical approach to estimate the substation equivalent load model parameters for supplying to oil-producing industrial region. The disadvantages of existing statistical approach are the low accuracy obtained for the parameter estimates, especially when using samples size is small. To eliminate this deficiency, the current measurement data archive from SCADA system of electrical parameters for 15 months was collected. For the purpose of verifying the obtained results of statistical processing of SCADA data, a full-scale experiment was carried out in relation to the studied substation. The article describes the statistical method used to process the current SCADA measurement data, the results of archived statistical processing and experimental SCADA data. The electrical load models' parameters received from the experimental studies results are of practical importance.

关键词： Load modeling Mathematical model Voltage measurement data models Clustering algorithms Biological system modeling Aggregates Load modeling power system power system study static load model ZIP model

来源：评论

学校读者我要写书评

暂无评论

Software Module Clustering Using the Hierarchical Clustering Combination Method 7

Software Module Clustering Using the Hierarchical Clustering...

引用

7th international conference on Cloud Computing and Big data Analytics, ICCCBDA 2022

作者： Xia, Hong Zhang, Yongkang Chen, Yanping Zhang, Hengshan Wang, Zhongmin Wang, Fengwei School of Computer Science and Technology Xi'An Univ. of P. and Telecom. Shaanxi Key Lab. of Network Data Analysis and Intelligent Processing Xi'an China Xi'An University of Posts and Telecommunications School of Computer Science and Technology Xi'an China ZTECorporation Shenzhen China

ISBN: (纸本)9781665487115

The scale of software applications has increased dramatically. Hierarchical clustering is a good method for modular recovery of software architecture. Because the different evaluation criteria of software types and clustering results, a single hierarchical clustering algorithm cannot integrate the number of clusters, arbitrary decision-making, recovery quality and other indicators on different evaluation standards, there is no comprehensive clustering algorithm selection method. We propose a hierarchical clustering combination algorithm and use principal component analysis to combine the results of multiple hierarchical clusters, the combined result retain the basic information of each clustering algorithm as much as possible. The experimental results show that compared with the single software hierarchical clustering method and other combined methods, the result of our method is closer to expert decomposition, and it has good performance on a variety of indicators. © 2022 IEEE.

关键词： Cloud computing Software architecture Clustering methods Software algorithms Decision making Clustering algorithms Big data

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：