Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954843
E. Ivannikova
This article proposes a scalable version of the Dependence Clustering algorithm which belongs to the class of spectral clustering methods. The method is implemented in Apache Spark using GraphX API primitives. Moreover, a fast approximate diffusion procedure that enables algorithms of spectral clustering type in Spark environment is introduced. In addition, the proposed algorithm is benchmarked against Spectral clustering. Results of applying the method to real-life data allow concluding that the implementation scales well, yet demonstrating good performance for densely connected graphs.
{"title":"Scalable implementation of dependence clustering in Apache Spark","authors":"E. Ivannikova","doi":"10.1109/EAIS.2017.7954843","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954843","url":null,"abstract":"This article proposes a scalable version of the Dependence Clustering algorithm which belongs to the class of spectral clustering methods. The method is implemented in Apache Spark using GraphX API primitives. Moreover, a fast approximate diffusion procedure that enables algorithms of spectral clustering type in Spark environment is introduced. In addition, the proposed algorithm is benchmarked against Spectral clustering. Results of applying the method to real-life data allow concluding that the implementation scales well, yet demonstrating good performance for densely connected graphs.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124098109","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 : 2017-05-01DOI: 10.1109/EAIS.2017.7954839
R. Precup, C. Dragos, Elena-Lorena Hedrea, Marian-Dan Rarinca, E. Petriu
This paper proposes evolving Takagi-Sugeno (T-S) fuzzy models that characterize the nonlinear dynamics phenomena occurring in the position of magnetic levitation systems. A state feedback control structure is first designed to stabilize the nonlinear process by linearization at certain operating points, and the evolving T-S fuzzy models are next derived for the stabilized closed-loop system. The rule bases and the parameters of the T-S fuzzy models are evolved by an incremental online identification algorithm (OIA). Real-time experiments are conducted in order to validate the evolving T-S fuzzy models that give the sphere position in magnetic levitation system laboratory equipment. The experimental results prove the very good performance of the T-S fuzzy models in terms of output responses and root mean square error values. The performance comparison with similar T-S fuzzy models evolved by another incremental OIA and three nature-inspired optimization algorithms is included.
{"title":"Evolving fuzzy models for the position control of magnetic levitation systems","authors":"R. Precup, C. Dragos, Elena-Lorena Hedrea, Marian-Dan Rarinca, E. Petriu","doi":"10.1109/EAIS.2017.7954839","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954839","url":null,"abstract":"This paper proposes evolving Takagi-Sugeno (T-S) fuzzy models that characterize the nonlinear dynamics phenomena occurring in the position of magnetic levitation systems. A state feedback control structure is first designed to stabilize the nonlinear process by linearization at certain operating points, and the evolving T-S fuzzy models are next derived for the stabilized closed-loop system. The rule bases and the parameters of the T-S fuzzy models are evolved by an incremental online identification algorithm (OIA). Real-time experiments are conducted in order to validate the evolving T-S fuzzy models that give the sphere position in magnetic levitation system laboratory equipment. The experimental results prove the very good performance of the T-S fuzzy models in terms of output responses and root mean square error values. The performance comparison with similar T-S fuzzy models evolved by another incremental OIA and three nature-inspired optimization algorithms is included.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127672199","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 : 2017-05-01DOI: 10.1109/EAIS.2017.7954844
Markus Thill, W. Konen, Thomas Bäck
An unresolved challenge for all kind of temporal data is the reliable anomaly detection, especially when adaptability is required in the case of non-stationary time series or when the nature of future anomalies is unknown or only vaguely defined. Most of the current anomaly detection algorithms follow the general idea to classify an anomaly as a significant deviation from the prediction. In this paper we present a comparative study where several online anomaly detection algorithms are compared on the large Yahoo Webscope S5 anomaly benchmark. We show that a relatively Simple Online Regression Anomaly Detector (SORAD) is quite successful compared to other anomaly detectors. We discuss the importance of several adaptive and online elements of the algorithm and their influence on the overall anomaly detection accuracy.
{"title":"Online anomaly detection on the webscope S5 dataset: A comparative study","authors":"Markus Thill, W. Konen, Thomas Bäck","doi":"10.1109/EAIS.2017.7954844","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954844","url":null,"abstract":"An unresolved challenge for all kind of temporal data is the reliable anomaly detection, especially when adaptability is required in the case of non-stationary time series or when the nature of future anomalies is unknown or only vaguely defined. Most of the current anomaly detection algorithms follow the general idea to classify an anomaly as a significant deviation from the prediction. In this paper we present a comparative study where several online anomaly detection algorithms are compared on the large Yahoo Webscope S5 anomaly benchmark. We show that a relatively Simple Online Regression Anomaly Detector (SORAD) is quite successful compared to other anomaly detectors. We discuss the importance of several adaptive and online elements of the algorithm and their influence on the overall anomaly detection accuracy.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124613212","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 : 2017-05-01DOI: 10.1109/EAIS.2017.7954821
Lucas Oliveira, Valter J. S. Leite, Jeferson Silva, F. Gomide
Exact feedback linearization is a powerful control approach, but has poor robustness properties. Lack of robustness yields inadequate performance and in practice may induce instability. This paper addresses an approach to improve the robustness of feedback linearized systems using a model reference adaptive control mechanism with an evolving participatory learning procedure. The granular evolving fuzzy robust feedback linearization approach is a way to robustly and efficiently control unknown nonlinear systems around given operating points. The result is a robust closed-loop control approach in which participatory learning is employed to estimate unknown nonlinearities online to cancel their effects in the feedback linearized system. A simulation example using a surge tank, a widely studied benchmark in the literature, shows that the performance of the granular evolving robust feedback linearization is higher than classic feedback linearization, fuzzy model reference, and indirect adaptive fuzzy control approaches.
{"title":"Granular evolving fuzzy robust feedback linearization","authors":"Lucas Oliveira, Valter J. S. Leite, Jeferson Silva, F. Gomide","doi":"10.1109/EAIS.2017.7954821","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954821","url":null,"abstract":"Exact feedback linearization is a powerful control approach, but has poor robustness properties. Lack of robustness yields inadequate performance and in practice may induce instability. This paper addresses an approach to improve the robustness of feedback linearized systems using a model reference adaptive control mechanism with an evolving participatory learning procedure. The granular evolving fuzzy robust feedback linearization approach is a way to robustly and efficiently control unknown nonlinear systems around given operating points. The result is a robust closed-loop control approach in which participatory learning is employed to estimate unknown nonlinearities online to cancel their effects in the feedback linearized system. A simulation example using a surge tank, a widely studied benchmark in the literature, shows that the performance of the granular evolving robust feedback linearization is higher than classic feedback linearization, fuzzy model reference, and indirect adaptive fuzzy control approaches.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114740958","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 : 2017-05-01DOI: 10.1109/EAIS.2017.7954841
R. A. Alphinas, H. Hansen, Torben Tambo
Non-adaptive proportional controllers suffer from the ability to handle a system disturbance leading to a large steady-state error and undesired transient behavior. On the other hand, they are easy to implement and tune. This article examines whether an adaptive controller based on the MIT and Lyapunov principle leads to a more robust and accurate regulation. Both controllers have been tested on a thermodynamic system exposed to a disturbance. The experiment shows that the adaptive controller handles the disturbance faster and more accurate.
{"title":"Comparison of conventional closed-loop controller with an adaptive controller for a disturbed thermodynamic system","authors":"R. A. Alphinas, H. Hansen, Torben Tambo","doi":"10.1109/EAIS.2017.7954841","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954841","url":null,"abstract":"Non-adaptive proportional controllers suffer from the ability to handle a system disturbance leading to a large steady-state error and undesired transient behavior. On the other hand, they are easy to implement and tune. This article examines whether an adaptive controller based on the MIT and Lyapunov principle leads to a more robust and accurate regulation. Both controllers have been tested on a thermodynamic system exposed to a disturbance. The experiment shows that the adaptive controller handles the disturbance faster and more accurate.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"48 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131805538","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 : 2017-05-01DOI: 10.1109/EAIS.2017.7954824
O. Senyukova, V. Gavrishchaka, Ksenia Tulnova
Diagnostics and monitoring in applied and clinical psychology is often based on subjective patient's questionnaires and observations. Lack of objective quantitative approaches could lead to biased conclusions and selection of sub-optimal therapies. However, established methods of modern psychophysiology indicate possibility of objective physiological measurement of certain psychological states and their dynamics. Nevertheless treatment personalization and optimization is very difficult task even in medicine, where many objective diagnostic tools are available. Previously we have proposed generic quantitative framework capable of discovering optimal combination of physiological indicators for early detection of emerging pathologies and efficient multi-expert characterization of complex and rare states. Ability of implicit encoding of great variety of patterns and regimes in training phase makes our system evolving in nature and capable of robust novelty detection without any formal online learning algorithms. Here we argue that the same approach could be also applicable to objective psychophysiological monitoring and fast discovery of effective personalized therapies in applied and clinical psychology. The web-based version of our system will be made available for researchers and psychology practitioners.
{"title":"Multi-expert evolving system for objective psychophysiological monitoring and fast discovery of effective personalized therapies","authors":"O. Senyukova, V. Gavrishchaka, Ksenia Tulnova","doi":"10.1109/EAIS.2017.7954824","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954824","url":null,"abstract":"Diagnostics and monitoring in applied and clinical psychology is often based on subjective patient's questionnaires and observations. Lack of objective quantitative approaches could lead to biased conclusions and selection of sub-optimal therapies. However, established methods of modern psychophysiology indicate possibility of objective physiological measurement of certain psychological states and their dynamics. Nevertheless treatment personalization and optimization is very difficult task even in medicine, where many objective diagnostic tools are available. Previously we have proposed generic quantitative framework capable of discovering optimal combination of physiological indicators for early detection of emerging pathologies and efficient multi-expert characterization of complex and rare states. Ability of implicit encoding of great variety of patterns and regimes in training phase makes our system evolving in nature and capable of robust novelty detection without any formal online learning algorithms. Here we argue that the same approach could be also applicable to objective psychophysiological monitoring and fast discovery of effective personalized therapies in applied and clinical psychology. The web-based version of our system will be made available for researchers and psychology practitioners.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"82 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133794061","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：