Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002990
Qing Liu, Jine Zhang
The study of differential-algebraic neural network is a new and fascinating field. In this paper, one kind of novel mathematical expression combining differential equation and algebraic equation is designed. Some sufficient conditions are presented via the mean value theorem of multi-valued differentials and the control theory of differential systems to ensure global asymptotic stability of complex-valued memristive differential-algebraic neural networks. Several criteria are given to assure that a unique equilibrium point of this model is existed, in addition, it is globally asymptotically stable via the properties of nonsingular M-matrices and definitions of stability. It is noteworthy that these conditions are an extension of existing works. Moreover, numerical simulations are given to test theoretical results.
{"title":"Evolution Properties of Complex-Valued Memristive Differential-Algebraic Neural Networks","authors":"Qing Liu, Jine Zhang","doi":"10.1109/SSCI44817.2019.9002990","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002990","url":null,"abstract":"The study of differential-algebraic neural network is a new and fascinating field. In this paper, one kind of novel mathematical expression combining differential equation and algebraic equation is designed. Some sufficient conditions are presented via the mean value theorem of multi-valued differentials and the control theory of differential systems to ensure global asymptotic stability of complex-valued memristive differential-algebraic neural networks. Several criteria are given to assure that a unique equilibrium point of this model is existed, in addition, it is globally asymptotically stable via the properties of nonsingular M-matrices and definitions of stability. It is noteworthy that these conditions are an extension of existing works. Moreover, numerical simulations are given to test theoretical results.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"1 1","pages":"1255-1262"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89937726","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003046
Fan Wang, Jianqi Shi, Xuan Tang, Jielong Guo, Peidong Liang, Yuanzhi Feng
Automatic identification for traffic signs is an important part of intelligent driving and traffic safety. Deep learning has already made a great achievement in traffic sign detection. However, the camera on a car may capture a low resolution and blurry image in certain environments, which make it inaccurate for traffic sign detection. Therefore, we propose a method based on image super-resolution reconstruction for improving the detection rate of traffic signs. Firstly, a low-resolution image is transformed by CNN-based super-resolution network into a high-resolution one. Then, to meet the requirements of on-line processing, we use the generated super-resolution image as input for the detection network with 16 filters in this layer. At last, we separately trained two CNNs for super-resolution reconstruction and traffic sign detection, which reduce the processing time. Experimental results demonstrate that our model can achieve better performance than the existing methods for traffic sign detection.
{"title":"CNN-based Super-resolution Reconstruction for Traffic Sign Detection","authors":"Fan Wang, Jianqi Shi, Xuan Tang, Jielong Guo, Peidong Liang, Yuanzhi Feng","doi":"10.1109/SSCI44817.2019.9003046","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003046","url":null,"abstract":"Automatic identification for traffic signs is an important part of intelligent driving and traffic safety. Deep learning has already made a great achievement in traffic sign detection. However, the camera on a car may capture a low resolution and blurry image in certain environments, which make it inaccurate for traffic sign detection. Therefore, we propose a method based on image super-resolution reconstruction for improving the detection rate of traffic signs. Firstly, a low-resolution image is transformed by CNN-based super-resolution network into a high-resolution one. Then, to meet the requirements of on-line processing, we use the generated super-resolution image as input for the detection network with 16 filters in this layer. At last, we separately trained two CNNs for super-resolution reconstruction and traffic sign detection, which reduce the processing time. Experimental results demonstrate that our model can achieve better performance than the existing methods for traffic sign detection.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"27 1","pages":"1208-1213"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90022447","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002677
Shengkun Xie, A. Lawniczak, Junlin Hao, Chong Gan
Dimension reduction of data generated from a complex simulation model is an important aspect, for the purpose of better understanding the behaviour of data, and it is often needed in many fields of study, including computer simulation and modelling. Also, improving data explainability is highly desirable for studying dynamics of complex simulation models, dynamics of which depends on many parameters, and has become an important aspect in machine learning and artificial intelligence. In this work, we initiate an approach, combining principal component analysis, K-means clustering and ANOVA-F test, in order to analyze the data from a designed simulation experiment. We propose a new method for optimal selection of numbers of clusters for data clustering. The proposed method is illustrated by an analysis of agent-based computer simulation. Our study has demonstrated the usefulness of the proposed method in both explainable data analytic and analysis of complex systems.
{"title":"Improving Data Explainability in Analysis of Designed Computer Simulation Experiments","authors":"Shengkun Xie, A. Lawniczak, Junlin Hao, Chong Gan","doi":"10.1109/SSCI44817.2019.9002677","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002677","url":null,"abstract":"Dimension reduction of data generated from a complex simulation model is an important aspect, for the purpose of better understanding the behaviour of data, and it is often needed in many fields of study, including computer simulation and modelling. Also, improving data explainability is highly desirable for studying dynamics of complex simulation models, dynamics of which depends on many parameters, and has become an important aspect in machine learning and artificial intelligence. In this work, we initiate an approach, combining principal component analysis, K-means clustering and ANOVA-F test, in order to analyze the data from a designed simulation experiment. We propose a new method for optimal selection of numbers of clusters for data clustering. The proposed method is illustrated by an analysis of agent-based computer simulation. Our study has demonstrated the usefulness of the proposed method in both explainable data analytic and analysis of complex systems.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"8 12 1","pages":"1486-1493"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89675357","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002964
Daich Azuma, Y. Fukuyama, Akihiro Oi, Toru Jintsugawa, H. Fujimoto
This paper proposes parallel multi-population improved brain storm optimization with differential evolution strategies (PMP-IBSODE) for state estimation in distribution systems (SEDS) using just in time (JIT) modeling and correntropy. SEDS is a function which estimates system conditions such as voltage and current everywhere in the distribution system using limited measurement data. When outliers, which are not true values, are measured at the measurement points, JIT modeling and correntropy can be effective. Moreover, application of evolutionary computation techniques is necessary for the SEDS considering of a nonlinear characteristic of an objective function caused by equipment in distribution systems. Various evolutionary computation techniques including IBSODE have been applied to the SEDS. However, speed-up of calculation and high quality estimation results are required because of penetration of renewable energies. An evolutionary computation technique using multi-population and parallel distributed computing is one of solutions for the challenges. The proposed method is verified to speed up computation time and obtain higher quality estimation results than conventional methods.
{"title":"Parallel Multi-population Improved Brain Storm Optimization with Differential Evolution strategies for State Estimation in Distribution Systems using Just in Time Modeling and Correntropy","authors":"Daich Azuma, Y. Fukuyama, Akihiro Oi, Toru Jintsugawa, H. Fujimoto","doi":"10.1109/SSCI44817.2019.9002964","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002964","url":null,"abstract":"This paper proposes parallel multi-population improved brain storm optimization with differential evolution strategies (PMP-IBSODE) for state estimation in distribution systems (SEDS) using just in time (JIT) modeling and correntropy. SEDS is a function which estimates system conditions such as voltage and current everywhere in the distribution system using limited measurement data. When outliers, which are not true values, are measured at the measurement points, JIT modeling and correntropy can be effective. Moreover, application of evolutionary computation techniques is necessary for the SEDS considering of a nonlinear characteristic of an objective function caused by equipment in distribution systems. Various evolutionary computation techniques including IBSODE have been applied to the SEDS. However, speed-up of calculation and high quality estimation results are required because of penetration of renewable energies. An evolutionary computation technique using multi-population and parallel distributed computing is one of solutions for the challenges. The proposed method is verified to speed up computation time and obtain higher quality estimation results than conventional methods.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"52 1","pages":"2714-2720"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89677936","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003156
Iko Nakari, Akinori Murata, Eiki Kitajima, Hiroyuki Sato, K. Takadama
This paper proposes the new Sleep Apnea Syndrome (SAS) detection method based on Random Forest (RF) by estimating WAKE stage (shallow sleep) and analyzing characteristics of biological vibration data at WAKE stage. In particular, the proposed method estimates the WAKE stage from the biological vibration data acquired by the mattress sensor, and detects SAS based on the differences in the distribution of contribution of each frequency to classify the WAKE stage. To investigate the effectiveness of the proposed method, in cooperation with medical institutions, we applied the proposed method to a total of 18 subjects (nine SAS patients and nine healthy subjects). The results derive the following implications: (1) SAS patients have WAKE with small biological vibrations, and the contribution of the corresponding low frequency components is high while the high frequency components, which is corresponded to large biological vibrations, is low contribution; (2) the proposed method could correctly detect SAS with 100% accuracy and non-SAS with 77.8% accuracy.
{"title":"Sleep Apnea Syndrome Detection based on Biological Vibration Data from Mattress Sensor","authors":"Iko Nakari, Akinori Murata, Eiki Kitajima, Hiroyuki Sato, K. Takadama","doi":"10.1109/SSCI44817.2019.9003156","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003156","url":null,"abstract":"This paper proposes the new Sleep Apnea Syndrome (SAS) detection method based on Random Forest (RF) by estimating WAKE stage (shallow sleep) and analyzing characteristics of biological vibration data at WAKE stage. In particular, the proposed method estimates the WAKE stage from the biological vibration data acquired by the mattress sensor, and detects SAS based on the differences in the distribution of contribution of each frequency to classify the WAKE stage. To investigate the effectiveness of the proposed method, in cooperation with medical institutions, we applied the proposed method to a total of 18 subjects (nine SAS patients and nine healthy subjects). The results derive the following implications: (1) SAS patients have WAKE with small biological vibrations, and the contribution of the corresponding low frequency components is high while the high frequency components, which is corresponded to large biological vibrations, is low contribution; (2) the proposed method could correctly detect SAS with 100% accuracy and non-SAS with 77.8% accuracy.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"50 1","pages":"550-556"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90082482","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003162
Haw-Lih Su, Jeff Cheng-Lung Lee, Mohammad Ibrahim Ahmad Ibrahim, Mohammed Alsafran, S. Al-Meer
Drug detection plays the key role in the enforcement of drug regulations. Various detection methods for them are thus developed. However, in most investigations, the most time-consuming and tedious part is not on the analysis itself but on the process to figure out what it is, i.e. to find out the possibility and to match the analytical results with the known information. Because of the limited amounts of the sample, an improper starting of the test could result in consuming of the evidence or unsuccessful analysis. Here we suggest a new process starting from NMR spectroscopy as NMR measurement is a non-destructive analysis, allowing other analysis methods after it. A spectra database screening system could help for quickly finding out the possible candidates and provide suggestions to confirm the results.
{"title":"Database Facilitated Screening with NMR Spectroscopy Analysis for Drug Detection","authors":"Haw-Lih Su, Jeff Cheng-Lung Lee, Mohammad Ibrahim Ahmad Ibrahim, Mohammed Alsafran, S. Al-Meer","doi":"10.1109/SSCI44817.2019.9003162","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003162","url":null,"abstract":"Drug detection plays the key role in the enforcement of drug regulations. Various detection methods for them are thus developed. However, in most investigations, the most time-consuming and tedious part is not on the analysis itself but on the process to figure out what it is, i.e. to find out the possibility and to match the analytical results with the known information. Because of the limited amounts of the sample, an improper starting of the test could result in consuming of the evidence or unsuccessful analysis. Here we suggest a new process starting from NMR spectroscopy as NMR measurement is a non-destructive analysis, allowing other analysis methods after it. A spectra database screening system could help for quickly finding out the possible candidates and provide suggestions to confirm the results.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"154 1","pages":"219-224"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90418584","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003003
Yonghong Huang, Joanna Negrete, Adam Wosotowsky, John Wagener, Eric Peterson, Armando Rodriguez, Celeste Fralick
From the fundamentals of the domain name system (DNS) system, to the websites we browse, the files we download, and emails we receive, every aspect of our online lives involves connections to internet resources. As a result, the Internet protocol (IP) Address is a pivotal component for risk assessment of online exchanges. Our goal in this study is to develop large- scale classification of malicious IPs that leverages cross-protocol telemetry to produce accurate and context-aware risk assessment. We developed an IP reputation system for generic IP addresses based on real-world data. We added interpretability to our machine learning solution to infer a malicious IP address. Our results show that the cross-protocol analysis achieves exceptional testing performance and is effective in real-world application to detect malicious IP addresses.
{"title":"Detect Malicious IP Addresses using Cross-Protocol Analysis","authors":"Yonghong Huang, Joanna Negrete, Adam Wosotowsky, John Wagener, Eric Peterson, Armando Rodriguez, Celeste Fralick","doi":"10.1109/SSCI44817.2019.9003003","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003003","url":null,"abstract":"From the fundamentals of the domain name system (DNS) system, to the websites we browse, the files we download, and emails we receive, every aspect of our online lives involves connections to internet resources. As a result, the Internet protocol (IP) Address is a pivotal component for risk assessment of online exchanges. Our goal in this study is to develop large- scale classification of malicious IPs that leverages cross-protocol telemetry to produce accurate and context-aware risk assessment. We developed an IP reputation system for generic IP addresses based on real-world data. We added interpretability to our machine learning solution to infer a malicious IP address. Our results show that the cross-protocol analysis achieves exceptional testing performance and is effective in real-world application to detect malicious IP addresses.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"19 1","pages":"664-672"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90515605","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003007
Yue Yuan, Yanjiang Wang, Xue Chen, Fu Wei
A central and fundamental issue in cognitive neuroscience is to comprehend the relationship between human brain functional and structural connectivity. Previous studies normally focus on the relationship by predicting functional connectivity from structural connectivity and show there is a cohesive correlation between the two types of networks. In this paper, we investigate the relation by revealing the true anatomical connections from the functional correlations using multi-layer neural networks, which is trained to learn the intrinsic mapping mechanism and recover some missed connections with diffusion magnetic resonance imaging (dMRI) tractography, particularly the cross-hemispheric homotopic connections. We execute the method to a dataset with 246 brain areas acquired from 147 subjects. The results show that around 65% of the average intrahemispheric structural connections are correctly inferred.
{"title":"Inferring Human Brain Structural Connectivity Based on Neural Networks","authors":"Yue Yuan, Yanjiang Wang, Xue Chen, Fu Wei","doi":"10.1109/SSCI44817.2019.9003007","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003007","url":null,"abstract":"A central and fundamental issue in cognitive neuroscience is to comprehend the relationship between human brain functional and structural connectivity. Previous studies normally focus on the relationship by predicting functional connectivity from structural connectivity and show there is a cohesive correlation between the two types of networks. In this paper, we investigate the relation by revealing the true anatomical connections from the functional correlations using multi-layer neural networks, which is trained to learn the intrinsic mapping mechanism and recover some missed connections with diffusion magnetic resonance imaging (dMRI) tractography, particularly the cross-hemispheric homotopic connections. We execute the method to a dataset with 246 brain areas acquired from 147 subjects. The results show that around 65% of the average intrahemispheric structural connections are correctly inferred.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"11 1","pages":"1585-1589"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89248171","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002777
Shubham Pateria, Budhitama Subagdja, A. Tan
In the complex multi-agent tasks, various agents must cooperate to distribute relevant subtasks among each other to achieve joint task objectives. An agent’s choice of the relevant subtask changes over time with the changes in the task environment state. Multi-agent Hierarchical Reinforcement Learning (MAHRL) provides an approach for learning to select the subtasks in response to the environment states, by using the joint task rewards to train various agents. When the joint task involves complex inter-agent dependencies, only a subset of agents might be capable of reaching the rewarding task states while other agents take precursory or intermediate roles. The delayed task reward might not be sufficient in such tasks to learn the coordinating policies for various agents. In this paper, we introduce a novel approach of MAHRL called Inter-Subtask Empowerment based Multi-agent Options (ISEMO) in which an Inter-Subtask Empowerment Reward (ISER) is given to an agent which enables the precondition(s) of other agents’ subtasks. ISER is given in addition to the domain task reward in order to improve the inter-agent coordination. ISEMO also incorporates options model that can learn parameterized subtask termination functions and relax the limitations posed by hand-crafted termination conditions. Experiments in a spatial Search and Rescue domain show that ISEMO can learn the subtask selection policies of various agents grounded in the inter-dependencies among the agents, as well as learn the subtask termination conditions, and perform better than the standard MAHRL technique.
{"title":"Multi-agent Reinforcement Learning in Spatial Domain Tasks using Inter Subtask Empowerment Rewards","authors":"Shubham Pateria, Budhitama Subagdja, A. Tan","doi":"10.1109/SSCI44817.2019.9002777","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002777","url":null,"abstract":"In the complex multi-agent tasks, various agents must cooperate to distribute relevant subtasks among each other to achieve joint task objectives. An agent’s choice of the relevant subtask changes over time with the changes in the task environment state. Multi-agent Hierarchical Reinforcement Learning (MAHRL) provides an approach for learning to select the subtasks in response to the environment states, by using the joint task rewards to train various agents. When the joint task involves complex inter-agent dependencies, only a subset of agents might be capable of reaching the rewarding task states while other agents take precursory or intermediate roles. The delayed task reward might not be sufficient in such tasks to learn the coordinating policies for various agents. In this paper, we introduce a novel approach of MAHRL called Inter-Subtask Empowerment based Multi-agent Options (ISEMO) in which an Inter-Subtask Empowerment Reward (ISER) is given to an agent which enables the precondition(s) of other agents’ subtasks. ISER is given in addition to the domain task reward in order to improve the inter-agent coordination. ISEMO also incorporates options model that can learn parameterized subtask termination functions and relax the limitations posed by hand-crafted termination conditions. Experiments in a spatial Search and Rescue domain show that ISEMO can learn the subtask selection policies of various agents grounded in the inter-dependencies among the agents, as well as learn the subtask termination conditions, and perform better than the standard MAHRL technique.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"669 1","pages":"86-93"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76855837","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002812
Mao Tan, Ji-Cheng Jin, Yongxin Su
In grid-connected multi-energy microgrid, fluctuation of renewable energy generation and coupling of multiple energy resources make the power load difficult to forecast accurately. In this paper, we focus on the short-term gateway load forecasting of grid-connected multi-energy microgrid. Consider spatial correlation between microgrid nodes, the information of multiple nodes, e.g., renewable energy access node, gas turbine access node and some critical load nodes, is utilized to implement information fusion forecasting. We propose an ensemble model that integrates GBRT, XGboost, Decison Tree and Seq2Seq to solve the problem. An IEEE33 bus system based simulation is conducted on an integrated platform with OpenDSS and Simulink. The experimental results show that the proposed approach outperforms several classical time series models with higher accuracy and better stability.
{"title":"An Ensemble Learning Approach for Short-Term Load Forecasting of Grid-Connected Multi-energy Microgrid","authors":"Mao Tan, Ji-Cheng Jin, Yongxin Su","doi":"10.1109/SSCI44817.2019.9002812","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002812","url":null,"abstract":"In grid-connected multi-energy microgrid, fluctuation of renewable energy generation and coupling of multiple energy resources make the power load difficult to forecast accurately. In this paper, we focus on the short-term gateway load forecasting of grid-connected multi-energy microgrid. Consider spatial correlation between microgrid nodes, the information of multiple nodes, e.g., renewable energy access node, gas turbine access node and some critical load nodes, is utilized to implement information fusion forecasting. We propose an ensemble model that integrates GBRT, XGboost, Decison Tree and Seq2Seq to solve the problem. An IEEE33 bus system based simulation is conducted on an integrated platform with OpenDSS and Simulink. The experimental results show that the proposed approach outperforms several classical time series models with higher accuracy and better stability.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"1 1","pages":"497-502"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75388484","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：