Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00028
Han Zhao, Xinyu Jin
The attention mechanism based models for computer vision and natural language processing are widely utilized, and action recognition in videos is no exception. In this paper, we develop a novel convolutional and recurrent network for action recognition which is "doubly deep" in spatial and temporal layers. First, in the feature extraction stage, we propose an improved p-non-local operations as a simple and effective component to capture long-distance dependencies with deep convolutional neural networks. Second, in the class prediction stage, we propose Fusion KeyLess Attention combining with the forward and backward bidirectional LSTM to learn the sequential nature of the data more efficiently and elegantly, which uses multi-epoch models fusion based on confusion matrix. Experiments on two heterogeneous datasets, HMDB51 and Hollywood2 show that our model has distinct advantages over traditional models also only utilizing RGB features for action recognition based on CNN and RNN.
{"title":"Human Action Recognition Based on Improved Fusion Attention CNN and RNN","authors":"Han Zhao, Xinyu Jin","doi":"10.1109/ICCIA49625.2020.00028","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00028","url":null,"abstract":"The attention mechanism based models for computer vision and natural language processing are widely utilized, and action recognition in videos is no exception. In this paper, we develop a novel convolutional and recurrent network for action recognition which is \"doubly deep\" in spatial and temporal layers. First, in the feature extraction stage, we propose an improved p-non-local operations as a simple and effective component to capture long-distance dependencies with deep convolutional neural networks. Second, in the class prediction stage, we propose Fusion KeyLess Attention combining with the forward and backward bidirectional LSTM to learn the sequential nature of the data more efficiently and elegantly, which uses multi-epoch models fusion based on confusion matrix. Experiments on two heterogeneous datasets, HMDB51 and Hollywood2 show that our model has distinct advantages over traditional models also only utilizing RGB features for action recognition based on CNN and RNN.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132988549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00010
Chenyu Wang, Zhongchen Miao, Yuefeng Lin, Hang Jiang, Jian Gao, Jidong Lu, Guangwei Shi
As rapid growth, Chinese financial derivative market is holding increasingly large proportions in entire domestic capital market as well as in global shares. To the nature of derivative instruments, plenty of market data features (such as prices and trading volumes) and off-market factors (such as financial news and policies) can directly impact on the price and risk in Chinese financial derivative markets, which is becoming more and more infeasible to model by using only traditional financial models and hand-crafted features. To alleviate the issue, in this paper we introduce some state-of-art deep neural network architectures and model two significant futures market price and risk indicators that are widely used by Chinese regulators, which are turn-over ratio (ratio of daily trading volumes and daily open interest volumes) and price basis (gap between futures price and corresponding spot product price). The extensive experimental results show that deep learning methods perform better prediction accuracy than traditional methods, among which convolutional LSTM achieves better results in most cases as it can capture local time-variant patterns. In addition, we also propose methods to exploit alternative off-market features (such as social media emotions and Baidu Search Index) with DNN models, which are proven beneficial to the price and risk prediction by rendering extra information than only market data.
{"title":"Modeling Price and Risk in Chinese Financial Derivative Market with Deep Neural Network Architectures","authors":"Chenyu Wang, Zhongchen Miao, Yuefeng Lin, Hang Jiang, Jian Gao, Jidong Lu, Guangwei Shi","doi":"10.1109/ICCIA49625.2020.00010","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00010","url":null,"abstract":"As rapid growth, Chinese financial derivative market is holding increasingly large proportions in entire domestic capital market as well as in global shares. To the nature of derivative instruments, plenty of market data features (such as prices and trading volumes) and off-market factors (such as financial news and policies) can directly impact on the price and risk in Chinese financial derivative markets, which is becoming more and more infeasible to model by using only traditional financial models and hand-crafted features. To alleviate the issue, in this paper we introduce some state-of-art deep neural network architectures and model two significant futures market price and risk indicators that are widely used by Chinese regulators, which are turn-over ratio (ratio of daily trading volumes and daily open interest volumes) and price basis (gap between futures price and corresponding spot product price). The extensive experimental results show that deep learning methods perform better prediction accuracy than traditional methods, among which convolutional LSTM achieves better results in most cases as it can capture local time-variant patterns. In addition, we also propose methods to exploit alternative off-market features (such as social media emotions and Baidu Search Index) with DNN models, which are proven beneficial to the price and risk prediction by rendering extra information than only market data.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"331 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116235462","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00009
U. Dampage, Yasiru Gunaratne, Ovindi Bandara, S. Silva, Vinushi Waraketiya
The knowledge of water inflow figures is paramount in decision making on the allocation for consumption for numerous purposes; irrigation, hydropower, domestic and industrial usage, and flood control. The understanding on how reservoir inflows are affected by different climatic and hydrological conditions is crucial to enable effective water management and downstream flood control. In this research, we propose a method using a Long Short -Term Memory (LSTM) Artificial Neural Network (ANN) to assist the aforesaid decision-making process. The Kotmale reservoir which is the uppermost reservoir in the Mahaweli reservoir complex in Sri Lanka was used as the test bed for this research. The ANN uses the runoff in the Kotmale reservoir catchment area and the effect of Sea Surface temperatures (SST) to make a forecast for seven days ahead. Three types of ANN are tested; Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN) and LSTM. The extensive field trials and validation endeavors found that the LSTM ANN provides superior performance in the aspects of accuracy and latency.
{"title":"Artificial Neural Network for Forecasting of Daily Reservoir Inflow: Case Study of the Kotmale Reservoir in Sri Lanka","authors":"U. Dampage, Yasiru Gunaratne, Ovindi Bandara, S. Silva, Vinushi Waraketiya","doi":"10.1109/ICCIA49625.2020.00009","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00009","url":null,"abstract":"The knowledge of water inflow figures is paramount in decision making on the allocation for consumption for numerous purposes; irrigation, hydropower, domestic and industrial usage, and flood control. The understanding on how reservoir inflows are affected by different climatic and hydrological conditions is crucial to enable effective water management and downstream flood control. In this research, we propose a method using a Long Short -Term Memory (LSTM) Artificial Neural Network (ANN) to assist the aforesaid decision-making process. The Kotmale reservoir which is the uppermost reservoir in the Mahaweli reservoir complex in Sri Lanka was used as the test bed for this research. The ANN uses the runoff in the Kotmale reservoir catchment area and the effect of Sea Surface temperatures (SST) to make a forecast for seven days ahead. Three types of ANN are tested; Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN) and LSTM. The extensive field trials and validation endeavors found that the LSTM ANN provides superior performance in the aspects of accuracy and latency.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"84 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124342615","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00011
Shengbing Xu, Zhiping Ouyang, Jiqiang Feng
For solving multi-objective optimization problems, this paper firstly combines a multi-objective evolutionary algorithm based on decomposition (MOEA/D) with good convergence and non-dominated sorting genetic algorithm II (NSGA-II) with good distribution to construct. Thus we propose a hybrid multi-objective optimization solving algorithm. Then, we consider that the population diversity needs to be improved while applying multi-objective particle swarm optimization (MOPSO) to solve the multi-objective optimization problems and an improved MOPSO algorithm is proposed. We give the distance function between the individual and the population, and the individual with the largest distance is selected as the global optimal individual to maintain population diversity. Finally, the simulation experiments are performed on the ZDTDTLZ test functions and track planning problems. The results indicate the better performance of the improved algorithms.
{"title":"An Improved Multi-objective Particle Swarm Optimization","authors":"Shengbing Xu, Zhiping Ouyang, Jiqiang Feng","doi":"10.1109/ICCIA49625.2020.00011","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00011","url":null,"abstract":"For solving multi-objective optimization problems, this paper firstly combines a multi-objective evolutionary algorithm based on decomposition (MOEA/D) with good convergence and non-dominated sorting genetic algorithm II (NSGA-II) with good distribution to construct. Thus we propose a hybrid multi-objective optimization solving algorithm. Then, we consider that the population diversity needs to be improved while applying multi-objective particle swarm optimization (MOPSO) to solve the multi-objective optimization problems and an improved MOPSO algorithm is proposed. We give the distance function between the individual and the population, and the individual with the largest distance is selected as the global optimal individual to maintain population diversity. Finally, the simulation experiments are performed on the ZDTDTLZ test functions and track planning problems. The results indicate the better performance of the improved algorithms.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121384965","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00031
Sinuo Fan, Fan-jie Meng
In order to make full use of the useful information in the massive video data and provide early warning of abnormal events, we propose a video prediction and abnormal detection algorithm. The algorithm designed a generation adversarial network with a single generator and dual discriminator to predict the video, and then performs anomaly detection on the basis of the video prediction frame. For training the model, various loss functions such as perceptual loss and optical flow loss are added to constrain the network. Extensive experiments on three publicly available datasets validate the effectiveness of our method in terms of various evaluation criteria.
{"title":"Video Prediction and Anomaly Detection Algorithm Based On Dual Discriminator","authors":"Sinuo Fan, Fan-jie Meng","doi":"10.1109/ICCIA49625.2020.00031","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00031","url":null,"abstract":"In order to make full use of the useful information in the massive video data and provide early warning of abnormal events, we propose a video prediction and abnormal detection algorithm. The algorithm designed a generation adversarial network with a single generator and dual discriminator to predict the video, and then performs anomaly detection on the basis of the video prediction frame. For training the model, various loss functions such as perceptual loss and optical flow loss are added to constrain the network. Extensive experiments on three publicly available datasets validate the effectiveness of our method in terms of various evaluation criteria.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115539309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00043
Yuqi Liu
The establishment of empathy is the premise and foundation for diverse innovative proposals and problem solutions. Virtual reality has provided a full range of depth and breadth for the establishment of empathy in many different types of fields due to its immersive, interactive, and imaginative characteristics. In this study, bibliometric analysis and VOSviewer software are used to cluster and visualize relevant 190 articles from the Web of Science core collection. The essay proposes a positioning of how to apply virtual reality on empathy based on two dimensions, from internal world to external world, and from business innovation to social innovation, by integrating each two of them, four application methods are summarized, which are meaning shaping, value creation, individual satisfaction, and self-realization. What’s more, using the bibliometric analysis result as a basis, the application landscape of virtual reality technology for establishing empathy has been constructed, including individual level, society level, and nature level, which reveals the existing and coming possibilities of using VR technology on building empathy in different fields. Last but not least, the paper has discussed the impact of virtual reality for empathy-building from five aspects, economy, politics, culture, society, and ecology. The efforts of this study reveal the VR tendency and have important reference significance for promoting the application of virtual reality technology in creating empathy and innovation in different fields.
同理心的建立是多元化创新建议和问题解决方案的前提和基础。虚拟现实以其沉浸式、互动性和想象力的特点，为在许多不同类型的领域建立共情提供了全方位的深度和广度。本研究采用文献计量学分析和VOSviewer软件对Web of Science核心馆藏的190篇相关文章进行聚类和可视化。本文从从内部世界到外部世界、从商业创新到社会创新两个维度对虚拟现实在共情中的应用进行了定位，并将这两个维度进行整合，总结出意义塑造、价值创造、个体满足和自我实现四种应用方法。并以文献计量分析结果为基础，构建了虚拟现实技术建立共情的应用景观，包括个体层面、社会层面和自然层面，揭示了虚拟现实技术在不同领域建立共情的现有和未来可能性。最后，本文从经济、政治、文化、社会和生态五个方面探讨了虚拟现实对移情建设的影响。本研究揭示了虚拟现实的趋势，对于推动虚拟现实技术在不同领域的移情和创新应用具有重要的参考意义。
{"title":"The application of virtual reality in empathy establishment: Foresee the future","authors":"Yuqi Liu","doi":"10.1109/ICCIA49625.2020.00043","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00043","url":null,"abstract":"The establishment of empathy is the premise and foundation for diverse innovative proposals and problem solutions. Virtual reality has provided a full range of depth and breadth for the establishment of empathy in many different types of fields due to its immersive, interactive, and imaginative characteristics. In this study, bibliometric analysis and VOSviewer software are used to cluster and visualize relevant 190 articles from the Web of Science core collection. The essay proposes a positioning of how to apply virtual reality on empathy based on two dimensions, from internal world to external world, and from business innovation to social innovation, by integrating each two of them, four application methods are summarized, which are meaning shaping, value creation, individual satisfaction, and self-realization. What’s more, using the bibliometric analysis result as a basis, the application landscape of virtual reality technology for establishing empathy has been constructed, including individual level, society level, and nature level, which reveals the existing and coming possibilities of using VR technology on building empathy in different fields. Last but not least, the paper has discussed the impact of virtual reality for empathy-building from five aspects, economy, politics, culture, society, and ecology. The efforts of this study reveal the VR tendency and have important reference significance for promoting the application of virtual reality technology in creating empathy and innovation in different fields.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115683110","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00046
Chengrong Wang, Xiaodong Zhang, Dian-Hui Chu
With the development of Cloud Computing, Internet of Things, and the advent of the era of Big Data, the types and scale of services are getting larger and larger, and the problem space of service composition is exploding. In order to measure the composite services quality of different combination schemes, this paper shows the calculation method of composite services QoS (Quality of Service), and improves the Ant Colony Algorithm by introducing Skyline calculation to further improve the efficiency of service composition and respond to user quickly. Finally, it is verified on the real QoS data set, and the feasibility and effectiveness of the method are proved through experiments.
随着云计算、物联网的发展和大数据时代的到来，服务的种类和规模越来越大，服务构成的问题空间呈爆炸式增长。为了衡量不同组合方案的组合服务质量，本文给出了组合服务QoS (quality of Service)的计算方法，并通过引入Skyline计算对蚁群算法进行改进，进一步提高了组合服务的效率，快速响应用户。最后，在真实的QoS数据集上进行了验证，通过实验验证了该方法的可行性和有效性。
{"title":"Research on Service Composition Optimization Method Based on Composite Services QoS","authors":"Chengrong Wang, Xiaodong Zhang, Dian-Hui Chu","doi":"10.1109/ICCIA49625.2020.00046","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00046","url":null,"abstract":"With the development of Cloud Computing, Internet of Things, and the advent of the era of Big Data, the types and scale of services are getting larger and larger, and the problem space of service composition is exploding. In order to measure the composite services quality of different combination schemes, this paper shows the calculation method of composite services QoS (Quality of Service), and improves the Ant Colony Algorithm by introducing Skyline calculation to further improve the efficiency of service composition and respond to user quickly. Finally, it is verified on the real QoS data set, and the feasibility and effectiveness of the method are proved through experiments.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130635363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/iccia49625.2020.00005
{"title":"ICCIA 2020 Opinion","authors":"","doi":"10.1109/iccia49625.2020.00005","DOIUrl":"https://doi.org/10.1109/iccia49625.2020.00005","url":null,"abstract":"","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129915976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00025
Oleksandr Vakhno, Long Ma
Differentiating between similar image inputs has always been one of the key tasks in machine learning. Inspired by the recent progress in the area of natural language processing, we introduce the image similarity learning model that considers the semantic scene similarity in its decision process. The architecture of the model is organized in a way to consider the similarity in the feature vectors of the images, as well as the semantic similarity in their generated captions, which are later combined to reach a more accurate result. We use Siamese-like network structure for parallel image processing and receiving the accurate results. Our model confirmed to improve the accuracy of a standard convolutional neural network and was validated on INRIA Holidays Dataset.
{"title":"A Hybrid Architecture for Semantic Image Similarity Learning","authors":"Oleksandr Vakhno, Long Ma","doi":"10.1109/ICCIA49625.2020.00025","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00025","url":null,"abstract":"Differentiating between similar image inputs has always been one of the key tasks in machine learning. Inspired by the recent progress in the area of natural language processing, we introduce the image similarity learning model that considers the semantic scene similarity in its decision process. The architecture of the model is organized in a way to consider the similarity in the feature vectors of the images, as well as the semantic similarity in their generated captions, which are later combined to reach a more accurate result. We use Siamese-like network structure for parallel image processing and receiving the accurate results. Our model confirmed to improve the accuracy of a standard convolutional neural network and was validated on INRIA Holidays Dataset.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126819799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00034
C. Yuan, L. Zhang
Spectral averagely-dense clustering is a clustering algorithm based on density, but it has the problem of being sensitive to the parameter ε. Aiming at the above problems, a spectral averagely-dense clustering based on dynamic shared nearest neighbors is put forward. Firstly, a similarity measures is constructed by combining self-tunning distance and shared nearest neighbors. Self-tunning distance can handle clusters of different density, and shared nearest neighbors can draw closer to the data in the same cluster and alienate the data in different clusters. Secondly, based on the sample distribution function, a method capable of self-adaptively determining the k-value of the shared nearest neighbors is proposed without setting the parameter k. Finally, the constructed similarity measure is used as the similarity measure of the fully connected graph. The ε-neighberhood graph of spectral averagely-dense clustering is replaced with the fully connected graph, which avoid setting the parameter ε. Through the experiments on artificial datasets and UCI datasets, the proposed algorithm is compared with the spectral averagelydense clustering and the standard spectral clustering. The experimental results show that the proposed algorithm not only avoids the problem of difficult selection of ε-neighberhood graph parameters, but also has better performance on the datasets.
{"title":"Spectral Averagely-dense Clustering Based on Dynamic Shared Nearest Neighbors","authors":"C. Yuan, L. Zhang","doi":"10.1109/ICCIA49625.2020.00034","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00034","url":null,"abstract":"Spectral averagely-dense clustering is a clustering algorithm based on density, but it has the problem of being sensitive to the parameter ε. Aiming at the above problems, a spectral averagely-dense clustering based on dynamic shared nearest neighbors is put forward. Firstly, a similarity measures is constructed by combining self-tunning distance and shared nearest neighbors. Self-tunning distance can handle clusters of different density, and shared nearest neighbors can draw closer to the data in the same cluster and alienate the data in different clusters. Secondly, based on the sample distribution function, a method capable of self-adaptively determining the k-value of the shared nearest neighbors is proposed without setting the parameter k. Finally, the constructed similarity measure is used as the similarity measure of the fully connected graph. The ε-neighberhood graph of spectral averagely-dense clustering is replaced with the fully connected graph, which avoid setting the parameter ε. Through the experiments on artificial datasets and UCI datasets, the proposed algorithm is compared with the spectral averagelydense clustering and the standard spectral clustering. The experimental results show that the proposed algorithm not only avoids the problem of difficult selection of ε-neighberhood graph parameters, but also has better performance on the datasets.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133672945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：