Pub Date : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642209
Jianfeng Wang, Zhirun Chen, Yichong Sun, Zhan Li, Dacheng Pei, B. Cai
This paper is concerned with issue of stability analysis and control synthesis for a class of switched LPV systems with stochastic mode switching. The stochastic switching process is further regarded as hidden semi-Markov switching by using concepts of semi-Markov kernel and emission probability. With the construction of a novel Lyapunov function that depends on accessible-mode, parameter as well as elapsed time, the mean-square stability (MSS) criteria are developed for the closed-loop system with the usage of designed controller. Numerically testable stability criteria are established in sense of MSS by introducing the upper bound sojourn time and some mathematical techniques. The theoretical results are testified by a numerical example to demonstrate the effectiveness of the designed stabilizing controller.
{"title":"Controller Design for Switched LPV Systems with Hidden semi-Markov Mode Switching","authors":"Jianfeng Wang, Zhirun Chen, Yichong Sun, Zhan Li, Dacheng Pei, B. Cai","doi":"10.1109/ICICIP53388.2021.9642209","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642209","url":null,"abstract":"This paper is concerned with issue of stability analysis and control synthesis for a class of switched LPV systems with stochastic mode switching. The stochastic switching process is further regarded as hidden semi-Markov switching by using concepts of semi-Markov kernel and emission probability. With the construction of a novel Lyapunov function that depends on accessible-mode, parameter as well as elapsed time, the mean-square stability (MSS) criteria are developed for the closed-loop system with the usage of designed controller. Numerically testable stability criteria are established in sense of MSS by introducing the upper bound sojourn time and some mathematical techniques. The theoretical results are testified by a numerical example to demonstrate the effectiveness of the designed stabilizing controller.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124364023","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642205
Yunong Zhang, Jianzhen Xiao
Via the paper, the ZE (Zhang equivalency) concept is firstly revisited, and then, on the basis of ZE, equation-type ZE and inequation-type ZE are further showed and discussed. By utilizing ZND (i.e., Zhang neural dynamics) method, we present the general expressions of equation-type ZE, and apply them in seven types of time-variant matrix situations. Then, we further discuss the inequation-type ZE and present the order-l (i.e, 01), order-2 (i.e, 02), and order-3 (i.e, 03) general expressions of inequation-type ZE. By utilizing these expressions, we hence propose the 01, 02, and 03 expressions of inequation-type ZE in six types of matrix situations for handling corresponding time-variant matrix inequation problems.
{"title":"Time-Variant and Time-Invariant Matrix Inequations of Zhang Equivalency Besides Matrix Equations","authors":"Yunong Zhang, Jianzhen Xiao","doi":"10.1109/ICICIP53388.2021.9642205","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642205","url":null,"abstract":"Via the paper, the ZE (Zhang equivalency) concept is firstly revisited, and then, on the basis of ZE, equation-type ZE and inequation-type ZE are further showed and discussed. By utilizing ZND (i.e., Zhang neural dynamics) method, we present the general expressions of equation-type ZE, and apply them in seven types of time-variant matrix situations. Then, we further discuss the inequation-type ZE and present the order-l (i.e, 01), order-2 (i.e, 02), and order-3 (i.e, 03) general expressions of inequation-type ZE. By utilizing these expressions, we hence propose the 01, 02, and 03 expressions of inequation-type ZE in six types of matrix situations for handling corresponding time-variant matrix inequation problems.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125526189","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642208
Wing W. Y. Ng, Zixin Zhou, Ting Wang
Faces contain abundant biological and sociological information. Inter-ethnicity identification using facial images has been intensively studied, while intra-ethnicity classification has received less attention. In this paper, we propose an Ensemble of Convolutional Autoencoders (E-CAE) model to attempt to distinguish Chinese, Japanese, and Korean faces and individuals from different regions of China. To accomplish this task, CJK and RoC datasets are built and E-CAE yields a classification accuracy of 80.69% on CJK dataset and 61.81% on RoC dataset. The experimental results demonstrate that our model outperforms existing methods for fine-grained ethnicity recognition in terms of accuracy and robustness. To our knowledge, this is the first work that performs fine-grained ethnicity recognition at the scale of provinces.
{"title":"Fine-Grained Facial Ethnicity Recognition Based on Dual Convolutional Autoencoders","authors":"Wing W. Y. Ng, Zixin Zhou, Ting Wang","doi":"10.1109/ICICIP53388.2021.9642208","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642208","url":null,"abstract":"Faces contain abundant biological and sociological information. Inter-ethnicity identification using facial images has been intensively studied, while intra-ethnicity classification has received less attention. In this paper, we propose an Ensemble of Convolutional Autoencoders (E-CAE) model to attempt to distinguish Chinese, Japanese, and Korean faces and individuals from different regions of China. To accomplish this task, CJK and RoC datasets are built and E-CAE yields a classification accuracy of 80.69% on CJK dataset and 61.81% on RoC dataset. The experimental results demonstrate that our model outperforms existing methods for fine-grained ethnicity recognition in terms of accuracy and robustness. To our knowledge, this is the first work that performs fine-grained ethnicity recognition at the scale of provinces.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129211397","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642204
Man-Fai Leung, Jun Wang, Hangjun Che
This paper is concerned with portfolio selection based on the Markowitz mean-variance framework using neurodynamic optimization. The portfolio optimization problem is formulated as a biconvex optimization problem. A two-timescale duplex neurodynamic approach is then applied for solving the profolio selection problem. The approach makes use of two recurrent neural networks (RNNs) which operate at different timescales for local search. A particle swarm optimization algorithm is employed to update the neuronal states of the two RNNs for global optima. Experimental results on four stock market datasets show the superior performance of the neurodynamic approach in terms of long-term expected returns.
{"title":"Another Two-Timescale Duplex Neurodynamic Approach to Portfolio Selection","authors":"Man-Fai Leung, Jun Wang, Hangjun Che","doi":"10.1109/ICICIP53388.2021.9642204","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642204","url":null,"abstract":"This paper is concerned with portfolio selection based on the Markowitz mean-variance framework using neurodynamic optimization. The portfolio optimization problem is formulated as a biconvex optimization problem. A two-timescale duplex neurodynamic approach is then applied for solving the profolio selection problem. The approach makes use of two recurrent neural networks (RNNs) which operate at different timescales for local search. A particle swarm optimization algorithm is employed to update the neuronal states of the two RNNs for global optima. Experimental results on four stock market datasets show the superior performance of the neurodynamic approach in terms of long-term expected returns.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"205 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133964973","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642194
Md Sakib Ullah Sourav, Xiaoyun Zhang, Huidong Wang
With the emergence of the COVID-19 pandemic, tackling mental health issues has become challenging too. A tendency has been observed in people spending more time on social media (SM) than usual and it has become the alternative source of interaction and news sharing. Previous research shows that intensive use of SM increases stress directly or indirectly. The aim of this study is to analyze the role of self-efficacy on information support from SM and COVID-19 depressions. To achieve this objective, a quantitative analysis was performed through an online questionnaire-based survey among SM users. The findings of this study prevail that with the help of effective information support from SM and through certain behavioral modifications with users’ high self-efficacy, COVID-19 stress might be lessened accordingly.
{"title":"Social Media as Information Support in Reducing COVID – 19 Depressions: Self-Efficacy as Mediator for Behavioral Modeling","authors":"Md Sakib Ullah Sourav, Xiaoyun Zhang, Huidong Wang","doi":"10.1109/ICICIP53388.2021.9642194","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642194","url":null,"abstract":"With the emergence of the COVID-19 pandemic, tackling mental health issues has become challenging too. A tendency has been observed in people spending more time on social media (SM) than usual and it has become the alternative source of interaction and news sharing. Previous research shows that intensive use of SM increases stress directly or indirectly. The aim of this study is to analyze the role of self-efficacy on information support from SM and COVID-19 depressions. To achieve this objective, a quantitative analysis was performed through an online questionnaire-based survey among SM users. The findings of this study prevail that with the help of effective information support from SM and through certain behavioral modifications with users’ high self-efficacy, COVID-19 stress might be lessened accordingly.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126983717","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}
This study evaluates Moderate resolution Imaging Spectroradiometer (MODIS) aboard terra Collection 6(C6) Dark Target (DT) Aerosol Optical Depth (AOD) products at 10 km (MOD04_10k) and 3 km (MOD04_3k) spatial resolution against ground-based AOD observations over Yuan Island in the North Yellow Sea. MOD04_3k retrievals show better performance, with larger number of collocations, larger percentage of observations falling within EE, smaller RMSE and MAE, especially for spring. MODIS comparability with ground-based data is not monotonic with Quality Assured Confidence(QAC) value, valid data(QAC >0) are the most accurate data set for both MOD04_10k and MOD04_3k. For seasonal analysis, trends of MOD04_10k and MOD04_3k are similar, MODIS AOD data for autumn perform best, whereas data quality for spring is the worst, which might be due to the dust aerosol effect. In summary, Terra/MODIS AOD data(QAC >0) at 3km are slightly more reliable than data at 10km over Yuan Island in the North Yellow Sea.
本研究将terra Collection 6(C6)上的中分辨率成像光谱仪(MODIS)在10 km (MOD04_10k)和3 km (MOD04_3k)空间分辨率下的暗目标(DT)气溶胶光学深度(AOD)产品与北黄海元岛的地面AOD观测结果进行了比较。MOD04_3k检索表现出更好的性能，搭配数量更多，在EE范围内的观测值百分比更大，RMSE和MAE更小，特别是在春季。MODIS与地面数据的可比性不是单调的，具有质量保证置信度(QAC)值，有效数据(QAC >0)是MOD04_10k和MOD04_3k最准确的数据集。MOD04_10k和MOD04_3k的季节变化趋势相似，秋季MODIS AOD数据表现最好，春季MODIS AOD数据质量最差，这可能与沙尘气溶胶的影响有关。综上所述，北黄海元岛3km波段的Terra/MODIS AOD数据(QAC >0)比10km波段的数据更可靠。
{"title":"Validation of Terra/MODIS 3KM and 10KM Aerosol Optical Depth Over Yuan Island in the North Yellow Sea","authors":"Yujuan Ma, Jianchao Fan, Yanlong Chen, Jianli Zhang","doi":"10.1109/ICICIP53388.2021.9642173","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642173","url":null,"abstract":"This study evaluates Moderate resolution Imaging Spectroradiometer (MODIS) aboard terra Collection 6(C6) Dark Target (DT) Aerosol Optical Depth (AOD) products at 10 km (MOD04_10k) and 3 km (MOD04_3k) spatial resolution against ground-based AOD observations over Yuan Island in the North Yellow Sea. MOD04_3k retrievals show better performance, with larger number of collocations, larger percentage of observations falling within EE, smaller RMSE and MAE, especially for spring. MODIS comparability with ground-based data is not monotonic with Quality Assured Confidence(QAC) value, valid data(QAC >0) are the most accurate data set for both MOD04_10k and MOD04_3k. For seasonal analysis, trends of MOD04_10k and MOD04_3k are similar, MODIS AOD data for autumn perform best, whereas data quality for spring is the worst, which might be due to the dust aerosol effect. In summary, Terra/MODIS AOD data(QAC >0) at 3km are slightly more reliable than data at 10km over Yuan Island in the North Yellow Sea.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115481731","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642166
Zanhao Liang, Xiaoqin Wang, Zhuo Chen, Xiaonan Luo
A method to predict how the human body temperature changes over time is presented in this paper. In this work, classic recurrent neural network and its two variants are used to predict body temperature, and their predictions are compared to evaluate performance. To collect the data used for training, videos which save the temperature value in the form of pixel are recorded with FLIR ONE PRO LT, a thermal sensor, and frames extracted from the video are converted into numbers with optical character recognition technology. To make our method more valuable, the temperatures at different condition, like motionless and walking, are predicted. Experiment results show that classic recurrent neural network outperforms its two variants, this may because GRU and LSTM have more parameters than classic RNN, when training data are not enough, GRU and LSTM are more likely to overfit than classic RNN.
本文提出了一种预测人体温度随时间变化的方法。在这项工作中，使用经典的递归神经网络及其两种变体来预测体温，并将其预测结果进行比较以评估性能。为了收集训练数据，使用热传感器FLIR ONE PRO LT记录以像素形式保存温度值的视频，并使用光学字符识别技术将视频中提取的帧转换为数字。为了使我们的方法更有价值，我们预测了不同状态下的温度，比如静止和行走。实验结果表明，经典递归神经网络优于其两种变体，这可能是因为GRU和LSTM比经典RNN具有更多的参数，当训练数据不足时，GRU和LSTM比经典RNN更容易过拟合。
{"title":"Body Temperature Prediction with Recurrent Neural Network and its Variants","authors":"Zanhao Liang, Xiaoqin Wang, Zhuo Chen, Xiaonan Luo","doi":"10.1109/ICICIP53388.2021.9642166","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642166","url":null,"abstract":"A method to predict how the human body temperature changes over time is presented in this paper. In this work, classic recurrent neural network and its two variants are used to predict body temperature, and their predictions are compared to evaluate performance. To collect the data used for training, videos which save the temperature value in the form of pixel are recorded with FLIR ONE PRO LT, a thermal sensor, and frames extracted from the video are converted into numbers with optical character recognition technology. To make our method more valuable, the temperatures at different condition, like motionless and walking, are predicted. Experiment results show that classic recurrent neural network outperforms its two variants, this may because GRU and LSTM have more parameters than classic RNN, when training data are not enough, GRU and LSTM are more likely to overfit than classic RNN.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"56 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128342208","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}
Aiming at the nonlinear and uncertain problems of upper limb exoskeleton rehabilitation robot (ULERR) during passive training, a sliding model controller based on radial basis neural network is designed in this paper. Firstly, a four-degree-of-freedom ULERR is designed for stroke patients in soft paralysis and spasticity, and a kinetic model was established. Secondly, RBF neural network is used to approximate the uncertainty caused by spastic disturbance of patients in the system. The weight in the neural network is replaced by a single parameter, and the adaptive algorithm is easy to adjust and has strong real-time performance. The asymptotic stability of the controller is verified by Lyapunov theorem. Finally, the desired training trajectory of the upper limb is obtained by a three-dimensional motion capture system, and the simulation experiments are carried out with Matlab software to prove that the proposed control method solves the chattering problem of traditional sliding mode control, to meet the control requirements of real-time rehabilitation training.
{"title":"Sliding Mode Control Algorithm of Upper Limb Exoskeleton Rehabilitation Robot Based on RBF Neural Network","authors":"Bangcheng Zhang, Shuai Liu, Ye Li, Zaixiang Pang, Yan-ling Hao, Xiyu Zhang","doi":"10.1109/ICICIP53388.2021.9642219","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642219","url":null,"abstract":"Aiming at the nonlinear and uncertain problems of upper limb exoskeleton rehabilitation robot (ULERR) during passive training, a sliding model controller based on radial basis neural network is designed in this paper. Firstly, a four-degree-of-freedom ULERR is designed for stroke patients in soft paralysis and spasticity, and a kinetic model was established. Secondly, RBF neural network is used to approximate the uncertainty caused by spastic disturbance of patients in the system. The weight in the neural network is replaced by a single parameter, and the adaptive algorithm is easy to adjust and has strong real-time performance. The asymptotic stability of the controller is verified by Lyapunov theorem. Finally, the desired training trajectory of the upper limb is obtained by a three-dimensional motion capture system, and the simulation experiments are carried out with Matlab software to prove that the proposed control method solves the chattering problem of traditional sliding mode control, to meet the control requirements of real-time rehabilitation training.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130872711","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642212
Hongzong Li, Jun Wang
The traveling salesman problem is known to be NP-hard and has numerous areas of applications. This paper proposes a collaborative neurodynamic optimization algorithm based on Boltzmann machines for solving the traveling salesman problem. A population of Boltzmann machines is employed for local search, and their initial states are repeatedly reinitialized by using the particle swarm optimization update rule for global repositioning. The efficacy of the proposed collaborative neurodynamic optimization algorithm is substantiated on four traveling salesman problem benchmark instances.
{"title":"A Collaborative Neurodynamic Optimization Algorithm Based on Boltzmann Machines for Solving the Traveling Salesman Problem","authors":"Hongzong Li, Jun Wang","doi":"10.1109/ICICIP53388.2021.9642212","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642212","url":null,"abstract":"The traveling salesman problem is known to be NP-hard and has numerous areas of applications. This paper proposes a collaborative neurodynamic optimization algorithm based on Boltzmann machines for solving the traveling salesman problem. A population of Boltzmann machines is employed for local search, and their initial states are repeatedly reinitialized by using the particle swarm optimization update rule for global repositioning. The efficacy of the proposed collaborative neurodynamic optimization algorithm is substantiated on four traveling salesman problem benchmark instances.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115057045","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642167
Ye Xingxing, Cai Guoyong, Wang Shunjie
With the continuous progress of mobile Internet technology and GPS positioning technology of mobile devices, Social Network and Location Based Services (LBS) are gradually converging to form Location Based Social Network (LBSN). POI (Point of Interest) recommendation systems face the problems of variable user interests, very sparse user and POI check-in matrices, and nonlinear interaction modeling. To address the above problems, a Graph-enhanced Attention Graph Neural Network model is proposed for POI recommendation (POI-GAGN in short). POI-GAGN mines user/POI node representations on user-POI interaction graph, user-user social interaction graph, and POI-POI association interaction graph through interaction node feature extraction module, learns POI attribute information representations through text feature extraction module, and extracts short-term preference representations of users through short-term preference extraction module. A graph-enhanced attention mechanism is designed to correlates node representations, attribute information representations of POI, and short-term preferences of users with each other to achieve better information fusion. Finally, we conduct sufficient experiments on two real datasets to prove that the recommendation effect of POI-GAGN is better than other current advanced POI recommendation methods, and POI-GAGN can better overcome the problems of data sparsity and cold start in recommendations.
随着移动互联网技术和移动设备GPS定位技术的不断进步，Social Network和Location Based Services (LBS)逐渐融合，形成Location Based Social Network (LBSN)。兴趣点(POI)推荐系统面临着用户兴趣变化、用户和兴趣点签入矩阵非常稀疏以及非线性交互建模等问题。为了解决上述问题，提出了一种用于POI推荐的图增强注意图神经网络模型(简称POI- gagn)。POI- gagn通过交互节点特征提取模块挖掘用户-POI交互图、用户-用户社交交互图、POI-POI关联交互图上的用户/POI节点表示，通过文本特征提取模块学习POI属性信息表示，通过短期偏好提取模块提取用户的短期偏好表示。设计了一种图增强关注机制，将节点表示、POI属性信息表示和用户短期偏好相互关联，实现更好的信息融合。最后，我们在两个真实数据集上进行了充分的实验，证明了POI- gagn的推荐效果优于目前其他先进的POI推荐方法，并且POI- gagn可以更好地克服推荐中的数据稀疏性和冷启动问题。
{"title":"POI Recommendation Based on Graph Enhanced Attention GNN","authors":"Ye Xingxing, Cai Guoyong, Wang Shunjie","doi":"10.1109/ICICIP53388.2021.9642167","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642167","url":null,"abstract":"With the continuous progress of mobile Internet technology and GPS positioning technology of mobile devices, Social Network and Location Based Services (LBS) are gradually converging to form Location Based Social Network (LBSN). POI (Point of Interest) recommendation systems face the problems of variable user interests, very sparse user and POI check-in matrices, and nonlinear interaction modeling. To address the above problems, a Graph-enhanced Attention Graph Neural Network model is proposed for POI recommendation (POI-GAGN in short). POI-GAGN mines user/POI node representations on user-POI interaction graph, user-user social interaction graph, and POI-POI association interaction graph through interaction node feature extraction module, learns POI attribute information representations through text feature extraction module, and extracts short-term preference representations of users through short-term preference extraction module. A graph-enhanced attention mechanism is designed to correlates node representations, attribute information representations of POI, and short-term preferences of users with each other to achieve better information fusion. Finally, we conduct sufficient experiments on two real datasets to prove that the recommendation effect of POI-GAGN is better than other current advanced POI recommendation methods, and POI-GAGN can better overcome the problems of data sparsity and cold start in recommendations.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132497253","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：