Pub Date : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170447
Jing Li, Yan Yang, Jie Hu
Multi-objective optimization problems (MOP) have not been completely solved due to their complexity. The evolutionary algorithm simulates the motor foraging mode of the biological group, which has certain advantages for solving the MOP, and can obtain the ε-pareto optimal solution. Particle swarm optimization (PSO) is well suitable for some evolutionary algorithms because of its fast convergence. Considering convergence, diversity and user preference information of multiple targets, we propose multi-objective particle swarm optimization algorithm with angle preference and three-archive sets (APTPSO). The validity of AP-TPSO is described by calculating the GD and SP values of the standard test functions.
{"title":"Multi-Objective Particle Swarm optimization Algorithm Based on Angle Preference and Three-Archive Sets","authors":"Jing Li, Yan Yang, Jie Hu","doi":"10.1109/ISKE47853.2019.9170447","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170447","url":null,"abstract":"Multi-objective optimization problems (MOP) have not been completely solved due to their complexity. The evolutionary algorithm simulates the motor foraging mode of the biological group, which has certain advantages for solving the MOP, and can obtain the ε-pareto optimal solution. Particle swarm optimization (PSO) is well suitable for some evolutionary algorithms because of its fast convergence. Considering convergence, diversity and user preference information of multiple targets, we propose multi-objective particle swarm optimization algorithm with angle preference and three-archive sets (APTPSO). The validity of AP-TPSO is described by calculating the GD and SP values of the standard test functions.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130512519","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170272
Yingdong Li, Jian Wang, Hongfei Lin, Shaowu Zhang, Zhihao Yang
Understanding the semantic and logical relationships between sentence pair is a difficult problem to be solved in natural language understanding tasks. Although the Enhanced Sequential Inference Model is simple in structure and performs well in SNLI, the limited capacity of the this model limits its further improvement of performance. Inspired by Res-Net, we propose the res-ESIM by introducing the residual connection into the ESIM model to expand the capacity of the ESIM while maintaining properties of simple structure and easy training. We explore the performance of res-ESIM with word embedding and the ability of using the contextual embedding to enhance its performance. In the experiments on SNLI, GloVe is used as word embedding for the convenience of comparing with published models. In the experiments on MultiNLI, the output of BERT-base based on different enhancement methods is used as contextual embedding. The experiment results on SNLI showed that our model achieves competitive performance in all models that haven’t employed additional contextualized word representations and the experiment results on MultiNLI showed that res-ESIM can have more performance improvement than the original ESIM when the information of embedding is enhanced.
{"title":"Residual Connected Enhanced Sequential Inference Model for Natural Language Inference","authors":"Yingdong Li, Jian Wang, Hongfei Lin, Shaowu Zhang, Zhihao Yang","doi":"10.1109/ISKE47853.2019.9170272","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170272","url":null,"abstract":"Understanding the semantic and logical relationships between sentence pair is a difficult problem to be solved in natural language understanding tasks. Although the Enhanced Sequential Inference Model is simple in structure and performs well in SNLI, the limited capacity of the this model limits its further improvement of performance. Inspired by Res-Net, we propose the res-ESIM by introducing the residual connection into the ESIM model to expand the capacity of the ESIM while maintaining properties of simple structure and easy training. We explore the performance of res-ESIM with word embedding and the ability of using the contextual embedding to enhance its performance. In the experiments on SNLI, GloVe is used as word embedding for the convenience of comparing with published models. In the experiments on MultiNLI, the output of BERT-base based on different enhancement methods is used as contextual embedding. The experiment results on SNLI showed that our model achieves competitive performance in all models that haven’t employed additional contextualized word representations and the experiment results on MultiNLI showed that res-ESIM can have more performance improvement than the original ESIM when the information of embedding is enhanced.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"532 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116581201","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170424
Ziping Gao, B. Peng, Tianrui Li
In this paper, we propose a method of fusing text edge semantics (FTES) for text semantic segmentation. FTES divides an image containing text into text semantic region, edge semantic region and background semantic region, where edge region is as a transitional region that splits text region from background region. At the same time, we design a text semantics segmentation network FTES-Net to detect arbitrarily shaped text regions in an images. We perform experiments on two public datasets containing a large number of non-linear text regions, and the results show that our proposed text region detection method can achieve comparable results.
{"title":"Text Detection by Fusing Text Edge Semantics in Arbitrary Shapes","authors":"Ziping Gao, B. Peng, Tianrui Li","doi":"10.1109/ISKE47853.2019.9170424","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170424","url":null,"abstract":"In this paper, we propose a method of fusing text edge semantics (FTES) for text semantic segmentation. FTES divides an image containing text into text semantic region, edge semantic region and background semantic region, where edge region is as a transitional region that splits text region from background region. At the same time, we design a text semantics segmentation network FTES-Net to detect arbitrarily shaped text regions in an images. We perform experiments on two public datasets containing a large number of non-linear text regions, and the results show that our proposed text region detection method can achieve comparable results.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"343 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116455085","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170442
Dongyan Wang, Jing Dong, D. Zhou, Xiaopeng Wei, Qiang Zhang
The performance of an emotion recognition system is determined by the quality of emotional features. In this paper, we propose a feature optimization algorithm based on image enhancement and present a convolutional recurrent model to realize emotional recognition of natural speech. For three-dimensional (3-D) log-Mel spectrum and 3-D spectrogram features, the fast gamma transformation with an adaptive threshold is adopted for feature enhancement to make full use of the dynamic characteristics of non-stationary speech signals. Meanwhile, the model combining Convolutional Neural Network (CNN) with the rectangular kernels and Long Short-Term Memory (LSTM) is used to complete speech emotion recognition tasks. Experiments are carried out on two public emotional datasets, and results demonstrate the good generalization ability and recognition performance of our proposed model.
{"title":"Speech Emotion Recognition Based on Image Enhancement","authors":"Dongyan Wang, Jing Dong, D. Zhou, Xiaopeng Wei, Qiang Zhang","doi":"10.1109/ISKE47853.2019.9170442","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170442","url":null,"abstract":"The performance of an emotion recognition system is determined by the quality of emotional features. In this paper, we propose a feature optimization algorithm based on image enhancement and present a convolutional recurrent model to realize emotional recognition of natural speech. For three-dimensional (3-D) log-Mel spectrum and 3-D spectrogram features, the fast gamma transformation with an adaptive threshold is adopted for feature enhancement to make full use of the dynamic characteristics of non-stationary speech signals. Meanwhile, the model combining Convolutional Neural Network (CNN) with the rectangular kernels and Long Short-Term Memory (LSTM) is used to complete speech emotion recognition tasks. Experiments are carried out on two public emotional datasets, and results demonstrate the good generalization ability and recognition performance of our proposed model.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129482831","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}
Convolutional neural networks (CNN) have the ability of self-adaptive learning features, which provides new ideas for fault diagnosis and analysis in the field of high-speed trains(HST). Combined with deep learning and wavelet transform, a diagnostic model for unknown compound faults based on capsule network is proposed. It is used to solve the problems of nonlinear of vibration signals and the difficulty of diagnosing unknown compound faults. Firstly, the collected vibration signal is converted into a spectrum map suitable for the network size and directly input into the convolution network layer for feature learning, which avoids the shortage of information loss caused by manual extraction of features. Secondly, the basic features detected by the convolutional layer are input into the capsule layer for combination and packaging of features. Finally, the fault condition is identified by the trained classifier. Experiments on different data sets collected in the laboratory simulation show that the diagnostic rate of this method for unknown compound faults is 90.31%, increasing by 7.94% to compared with the existing methods. Experiments were carried out using different types of unknown compound faults, and the generalization ability and robustness of the proposed model were verified.
{"title":"Unknown Compound Faults Diagnosis of High Speed Train Based on Capsule Network","authors":"Yingjun Zhang, Yongquan Jiang, Yan Yang, Yuxiao Gou, Weihua Zhang, Jinxiong Chen","doi":"10.1109/ISKE47853.2019.9170327","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170327","url":null,"abstract":"Convolutional neural networks (CNN) have the ability of self-adaptive learning features, which provides new ideas for fault diagnosis and analysis in the field of high-speed trains(HST). Combined with deep learning and wavelet transform, a diagnostic model for unknown compound faults based on capsule network is proposed. It is used to solve the problems of nonlinear of vibration signals and the difficulty of diagnosing unknown compound faults. Firstly, the collected vibration signal is converted into a spectrum map suitable for the network size and directly input into the convolution network layer for feature learning, which avoids the shortage of information loss caused by manual extraction of features. Secondly, the basic features detected by the convolutional layer are input into the capsule layer for combination and packaging of features. Finally, the fault condition is identified by the trained classifier. Experiments on different data sets collected in the laboratory simulation show that the diagnostic rate of this method for unknown compound faults is 90.31%, increasing by 7.94% to compared with the existing methods. Experiments were carried out using different types of unknown compound faults, and the generalization ability and robustness of the proposed model were verified.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133913860","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170307
DanChen Wang, Xiaosong Zhang, Yang Xu, H. Song
The security of data circulation is the core technology of data fusion and sharing service. The paper proposes multi-party circulation mechanism of the trusted data using to MPI communication. In order to achieve the trusted computing, this study proposes the computing service platform based on SMPC, which encapsulates the operation of sensitive data such as encryption key, password, user data, and etc.by trusted hardware using the security extension of Intel SGX. Meanwhile, aiming at these problems of semantic security and efficient processing ability, we chooses ElGamal homomorphic encryption system. In additional, SGX is extended to the remote authentication mechanism. System can support the deployment of hybrid cloud mode. Thus, the data security circulation can be satisfied. Compared to other methods, it has the advantage of model security and efficient communication.
{"title":"A Secure Multi-Party Computing System Based on SGX Technology for Trusted Data Circulation","authors":"DanChen Wang, Xiaosong Zhang, Yang Xu, H. Song","doi":"10.1109/ISKE47853.2019.9170307","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170307","url":null,"abstract":"The security of data circulation is the core technology of data fusion and sharing service. The paper proposes multi-party circulation mechanism of the trusted data using to MPI communication. In order to achieve the trusted computing, this study proposes the computing service platform based on SMPC, which encapsulates the operation of sensitive data such as encryption key, password, user data, and etc.by trusted hardware using the security extension of Intel SGX. Meanwhile, aiming at these problems of semantic security and efficient processing ability, we chooses ElGamal homomorphic encryption system. In additional, SGX is extended to the remote authentication mechanism. System can support the deployment of hybrid cloud mode. Thus, the data security circulation can be satisfied. Compared to other methods, it has the advantage of model security and efficient communication.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130779974","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170362
Jihong Zhu, J. Pei
Pruning can reduce the size of the model without reducing its performance. After pruning, the model can run in a small terminal flexibly. This paper proposes a new filter pruning method that uses soft filter pruning via a structural similarity index(FPSSI) to compress and prune the network. FPSSI uses the structural similarity index to measuring the difference between different filters, the filters with similar structures are pruned to achieve the purpose of compressing the Deep Convolutional Neural Networks(DNN) model. Compared to the norm-based approach to remove ”relatively low” importance filters, the proposed method takes into account the structure between the filters. When applied to the different classification benchmarks, our method validates its usefulness and advantages. In CIFAR10, the ResNet network uses the SFP-SSIM method to reduce 52% of FLOPs and has better accuracy.
{"title":"Filter Pruning via Structural Similarity Index for Deep Convolutional Neural Networks Acceleration","authors":"Jihong Zhu, J. Pei","doi":"10.1109/ISKE47853.2019.9170362","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170362","url":null,"abstract":"Pruning can reduce the size of the model without reducing its performance. After pruning, the model can run in a small terminal flexibly. This paper proposes a new filter pruning method that uses soft filter pruning via a structural similarity index(FPSSI) to compress and prune the network. FPSSI uses the structural similarity index to measuring the difference between different filters, the filters with similar structures are pruned to achieve the purpose of compressing the Deep Convolutional Neural Networks(DNN) model. Compared to the norm-based approach to remove ”relatively low” importance filters, the proposed method takes into account the structure between the filters. When applied to the different classification benchmarks, our method validates its usefulness and advantages. In CIFAR10, the ResNet network uses the SFP-SSIM method to reduce 52% of FLOPs and has better accuracy.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131261619","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170368
Yu Feng, Ming Gao, Zehui Zhang
How to increase the performance of Web service classification is one of the hot topics in current service classification research. Web service classification approach based on traditional machine learning and deep learning algorithm are greatly influenced by the data sparsity and bad data scalability, the increase of the amount of data will affect the classification performance. At the same time, the classification result is not accurate due to the change of quality of service (QoS) without considering the time factor. Aiming at the above-mentioned problems, this paper proposes a Web service quality classification method based on convolutional neural network algorithm. The main contributions of this paper are as follows∶ (l) Based on the collaborative filtering algorithm of service, web service recommendation is made from the similarity between services themselves by considering QoS parameters. The time factor of service is considered in the classification process, achieving higher classification performance. (2) A service quality classification method based on VGG-16 algorithm is proposed. In this method, GlobalAveragePooling2D replace the full connectivity layer of the CNN, which reduces the network parameter overabundance caused by too many fully connected layers. Also, unlike the full connectivity layer, which requires a lot of training and tuning parameters, GlobalAveragePooling2D reduces spatial parameters, and its local connections, weight sharing, and pooling operations have fewer connections and parameters, making it easier to train. In this paper, the optimized network and the typical machine learning algorithm and deep convolutional network are tested on the WS-DREAM data set. Experiments show that compared with other classifiers, the CNN presented in this paper has the highest balance accuracy score in experiments. The classification accuracy of the optimized CNN classifier is 98.88% and the balance accuracy score is 99.27%.
{"title":"Web Service QoS Classification Based on Optimized Convolutional Neural Network","authors":"Yu Feng, Ming Gao, Zehui Zhang","doi":"10.1109/ISKE47853.2019.9170368","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170368","url":null,"abstract":"How to increase the performance of Web service classification is one of the hot topics in current service classification research. Web service classification approach based on traditional machine learning and deep learning algorithm are greatly influenced by the data sparsity and bad data scalability, the increase of the amount of data will affect the classification performance. At the same time, the classification result is not accurate due to the change of quality of service (QoS) without considering the time factor. Aiming at the above-mentioned problems, this paper proposes a Web service quality classification method based on convolutional neural network algorithm. The main contributions of this paper are as follows∶ (l) Based on the collaborative filtering algorithm of service, web service recommendation is made from the similarity between services themselves by considering QoS parameters. The time factor of service is considered in the classification process, achieving higher classification performance. (2) A service quality classification method based on VGG-16 algorithm is proposed. In this method, GlobalAveragePooling2D replace the full connectivity layer of the CNN, which reduces the network parameter overabundance caused by too many fully connected layers. Also, unlike the full connectivity layer, which requires a lot of training and tuning parameters, GlobalAveragePooling2D reduces spatial parameters, and its local connections, weight sharing, and pooling operations have fewer connections and parameters, making it easier to train. In this paper, the optimized network and the typical machine learning algorithm and deep convolutional network are tested on the WS-DREAM data set. Experiments show that compared with other classifiers, the CNN presented in this paper has the highest balance accuracy score in experiments. The classification accuracy of the optimized CNN classifier is 98.88% and the balance accuracy score is 99.27%.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129449700","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170348
Xiongtao Zhang, Xingguang Pan, Shitong Wang
At present, crowdsourcing, as a distributed solution, provides an effective and cheap solution for solving large tasks. However, due to the difference of workers’ knowledge and skill, and the existence of fraudsters, the labels quality of crowdsourcing can’t be effectively controlled and guaranteed. This paper proposes a novel label quality improvement method based on ensemble TSK fuzzy classifier with high interpretability, i.e., EW-TSK-CS. Each subclassifier TSKnoise-FC is an improved zero-order TSK fuzzy classifier which is trained by noisy label training data and is more robust. The objective function of each fuzzy sub-classifier has considered the existence of label noise, and the fuzzy subclassifier has the ability to deal with uncertain data. All the subclassifier integrated together by augmenting the original noisy-free validation data space with the output of each subclassifier in an incremental way. The augmented validation data is conducted by running the classical FCM clustering methods on the augmented validation data and using KNN to obtain the dictionary data. The label noise correction mechanism is based on the dictionary data. The experimental results on datasets Adult, chess and waveform3 show that this method can effectively improve the label quality of crowdsourcing compared with tradition label noise robustness methods, ensemble methods, and classical TSK fuzzy classifiers.
{"title":"Label Quality Improvement in Crowdsourcing with Ensemble TSK Fuzzy Classifier","authors":"Xiongtao Zhang, Xingguang Pan, Shitong Wang","doi":"10.1109/ISKE47853.2019.9170348","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170348","url":null,"abstract":"At present, crowdsourcing, as a distributed solution, provides an effective and cheap solution for solving large tasks. However, due to the difference of workers’ knowledge and skill, and the existence of fraudsters, the labels quality of crowdsourcing can’t be effectively controlled and guaranteed. This paper proposes a novel label quality improvement method based on ensemble TSK fuzzy classifier with high interpretability, i.e., EW-TSK-CS. Each subclassifier TSKnoise-FC is an improved zero-order TSK fuzzy classifier which is trained by noisy label training data and is more robust. The objective function of each fuzzy sub-classifier has considered the existence of label noise, and the fuzzy subclassifier has the ability to deal with uncertain data. All the subclassifier integrated together by augmenting the original noisy-free validation data space with the output of each subclassifier in an incremental way. The augmented validation data is conducted by running the classical FCM clustering methods on the augmented validation data and using KNN to obtain the dictionary data. The label noise correction mechanism is based on the dictionary data. The experimental results on datasets Adult, chess and waveform3 show that this method can effectively improve the label quality of crowdsourcing compared with tradition label noise robustness methods, ensemble methods, and classical TSK fuzzy classifiers.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133423787","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 : 2019-11-01DOI: 10.1109/ISKE47853.2019.9170204
G. Khan, Jie Hu, Tianrui Li, Bassoma Diallo, Qianqian Huang
In recent years, datasets which exist in present world are comprising of various representations of the data or in multiview environment, which frequently give the important data to each other. Multi-view clustering based on Non-negative matrix factorization (NMF) has turned to be a very hot direction of research in the field of Pattern Reognition, Machine Learning (ML), and data mining. and data mining due to unsupervised confuse information of Numerous Views. The main problem of employing NMF to multi-view clustering is how to define the factorizations to give significant and commensurate clustering solutions. Specially, multi-view clustering based NMF has achieved extensive attention due to its dimensionality reduction property. Existing methods based on NMF barely produced meaningful clustering solution from heterogeneous numerous views due to their complementary behaviors. To address this issue, we design a innovative NMF technique based Multiview clustering approach, which gives the more meaningful and compatible clustering solution over Numerous Views. The main outcome of the work, is to a design combined NMF method with view weight and constraint co-efficient which will bring the clustering solution to a common point for each view. The effectiveness of propose method is validated by conducting the experiments on real-world datasets.
{"title":"Weighted Multi-View Data Clustering via Joint Non-Negative Matrix Factorization","authors":"G. Khan, Jie Hu, Tianrui Li, Bassoma Diallo, Qianqian Huang","doi":"10.1109/ISKE47853.2019.9170204","DOIUrl":"https://doi.org/10.1109/ISKE47853.2019.9170204","url":null,"abstract":"In recent years, datasets which exist in present world are comprising of various representations of the data or in multiview environment, which frequently give the important data to each other. Multi-view clustering based on Non-negative matrix factorization (NMF) has turned to be a very hot direction of research in the field of Pattern Reognition, Machine Learning (ML), and data mining. and data mining due to unsupervised confuse information of Numerous Views. The main problem of employing NMF to multi-view clustering is how to define the factorizations to give significant and commensurate clustering solutions. Specially, multi-view clustering based NMF has achieved extensive attention due to its dimensionality reduction property. Existing methods based on NMF barely produced meaningful clustering solution from heterogeneous numerous views due to their complementary behaviors. To address this issue, we design a innovative NMF technique based Multiview clustering approach, which gives the more meaningful and compatible clustering solution over Numerous Views. The main outcome of the work, is to a design combined NMF method with view weight and constraint co-efficient which will bring the clustering solution to a common point for each view. The effectiveness of propose method is validated by conducting the experiments on real-world datasets.","PeriodicalId":399084,"journal":{"name":"2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE)","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133430208","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：