Jiao Liu, Guoyou Shi, Kaige Zhu, Jiahui Shi, Yuchuang Wang
Aiming at the problems that the current decision-making model of ship collision avoidance does not consider International Regulations for Preventing Collisions at Sea (COLREGS), ship maneuverability, and the need for a lot of training time, combined with the advantages of reinforcement learning and imitation learning, a ship intelligent collision avoidance decision-making model based on Generic Adversary Imitation Learning (GAIL) is proposed: Firstly, the collision avoidance data in Automatic Information System (AIS) data is extracted as expert data; Secondly, in the generator part, the environment model is established based on Mathematical Model Group (MMG) and S-57 chart rendering, and the state space, behaviour space and reward function of reinforcement learning are constructed. The deep deterministic policy gradient (DDPG) is used to interact with the environment model to generate ship trajectory data. At the same time, the generator can constantly learn expert data; Finally, a discriminator can distinguish the expert data from the data generated by the generator is constructed and trained. The model training is completed when the discriminator cannot distinguish the two. In order to verify the performance of the model, AIS data near the South China Sea is used to process and extract collision avoidance decision data, and a ship intelligent collision avoidance decision model based on GAIL is established. After the model converges, the final generated data is compared with the expert data. The experimental results verify that the model proposed in this paper can reproduce the expert collision avoidance trajectory and is a practical decision model of ship collision avoidance.
{"title":"Decision Model of Ship Intelligent Collision Avoidance Based on Automatic Information System Data and Generic Adversary Imitation Learning-Deep Deterministic Policy Gradient","authors":"Jiao Liu, Guoyou Shi, Kaige Zhu, Jiahui Shi, Yuchuang Wang","doi":"10.1145/3583788.3583790","DOIUrl":"https://doi.org/10.1145/3583788.3583790","url":null,"abstract":"Aiming at the problems that the current decision-making model of ship collision avoidance does not consider International Regulations for Preventing Collisions at Sea (COLREGS), ship maneuverability, and the need for a lot of training time, combined with the advantages of reinforcement learning and imitation learning, a ship intelligent collision avoidance decision-making model based on Generic Adversary Imitation Learning (GAIL) is proposed: Firstly, the collision avoidance data in Automatic Information System (AIS) data is extracted as expert data; Secondly, in the generator part, the environment model is established based on Mathematical Model Group (MMG) and S-57 chart rendering, and the state space, behaviour space and reward function of reinforcement learning are constructed. The deep deterministic policy gradient (DDPG) is used to interact with the environment model to generate ship trajectory data. At the same time, the generator can constantly learn expert data; Finally, a discriminator can distinguish the expert data from the data generated by the generator is constructed and trained. The model training is completed when the discriminator cannot distinguish the two. In order to verify the performance of the model, AIS data near the South China Sea is used to process and extract collision avoidance decision data, and a ship intelligent collision avoidance decision model based on GAIL is established. After the model converges, the final generated data is compared with the expert data. The experimental results verify that the model proposed in this paper can reproduce the expert collision avoidance trajectory and is a practical decision model of ship collision avoidance.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114453801","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}
The classification accuracy of a multi-layer Perceptron Neural Networks depends on the selection of its parameters such the connection weights and biases. Generating an optimal value of these parameters requires a suitable algorithm to train the multilayer perceptron neural networks. This paper presents swam based Grasshopper optimization algorithm that optimizes the connection weights and biases of Multilayer Perceptron Neural Network. Grasshopper optimization algorithm is a swarm-based metaheuristic algorithm applied for accurate learning of Multilayer Perceptron Neural Networks. The proposed Multilayer Layer Perceptron Neural Networks based on the Grasshopper Optimization Algorithm was validated using a Genetic algorithm and Backpropagation algorithm this algorithm has proved to perform satisfactorily performance by escaping local optimal and its fast convergence.
{"title":"Improved Multilayer Perceptron Neural Networks Weights and Biases Based on The Grasshopper optimization Algorithm to Predict Student Performance on Ambient Learning","authors":"Mercy K. Michira, R. Rimiru, W. Mwangi","doi":"10.1145/3583788.3583797","DOIUrl":"https://doi.org/10.1145/3583788.3583797","url":null,"abstract":"The classification accuracy of a multi-layer Perceptron Neural Networks depends on the selection of its parameters such the connection weights and biases. Generating an optimal value of these parameters requires a suitable algorithm to train the multilayer perceptron neural networks. This paper presents swam based Grasshopper optimization algorithm that optimizes the connection weights and biases of Multilayer Perceptron Neural Network. Grasshopper optimization algorithm is a swarm-based metaheuristic algorithm applied for accurate learning of Multilayer Perceptron Neural Networks. The proposed Multilayer Layer Perceptron Neural Networks based on the Grasshopper Optimization Algorithm was validated using a Genetic algorithm and Backpropagation algorithm this algorithm has proved to perform satisfactorily performance by escaping local optimal and its fast convergence.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133647727","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}
Due to the excessive number of objective functions in DNA coding problem, there are dominant impedance between solutions which makes it difficult to evaluate the solutions and the algorithm is hard to converge. And traditional multi-objective evolutionary algorithms tend to fall into premature convergence when dealing with DNA coding problems. We proposed an Improved Nondominated Sorting Genetic Algorithm II with Constraint (ICNSAG-II) to deal with these problem. Firstly, the DNA coding problem and its 6 coding constraints are introduced. Secondly, the constraint function and Block operator are used to reduce the dimensionality of the DNA coding problem, so that the objective function is reduced to two, which make it easy to optimize using multi-objective evolutionary algorithms. Finally, by comparing with the sequences generated by the comparative algorithm, it was verified that the DNA sequences generated by ICNSGA-II have good chemical stability and are able to prevent the of unexpected secondary structures and non-specific hybridization reactions.
{"title":"A DNA Coding Design based on Multi-objective Evolutionary Algorithm with Constraint","authors":"Hengyu Duan, Kai Zhang, Xinbo Zhang","doi":"10.1145/3583788.3583794","DOIUrl":"https://doi.org/10.1145/3583788.3583794","url":null,"abstract":"Due to the excessive number of objective functions in DNA coding problem, there are dominant impedance between solutions which makes it difficult to evaluate the solutions and the algorithm is hard to converge. And traditional multi-objective evolutionary algorithms tend to fall into premature convergence when dealing with DNA coding problems. We proposed an Improved Nondominated Sorting Genetic Algorithm II with Constraint (ICNSAG-II) to deal with these problem. Firstly, the DNA coding problem and its 6 coding constraints are introduced. Secondly, the constraint function and Block operator are used to reduce the dimensionality of the DNA coding problem, so that the objective function is reduced to two, which make it easy to optimize using multi-objective evolutionary algorithms. Finally, by comparing with the sequences generated by the comparative algorithm, it was verified that the DNA sequences generated by ICNSGA-II have good chemical stability and are able to prevent the of unexpected secondary structures and non-specific hybridization reactions.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131381076","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}
The factor score method is a commonly used method to solve the problem of location, but the implementation process is greatly influenced by the subjective evaluation of experts, and the calculation results lack objectivity and accuracy. In order to overcome the above-mentioned shortcomings, this paper proposes a fuzzy multi-criteria decision-making model based on linguistic distribution evaluation method, entropy weight method and TODIM (interactive multi-criteria decision-making) method for location research. Firstly, this paper uses the linguistic distribution evaluation method to obtain the expert evaluation information, then uses the entropy weight method to calculate the weight of each influencing factor, and finally uses the TODIM method to determine the optimal location.Through case comparison and analysis, it is found that the model can effectively reduce the uncertainty of expert evaluation and improve the reliability of location results.
{"title":"Research on Location Problem Based on Fuzzy Multi-criteria Decision Method","authors":"Jinjin Ge, Mingshun Song, Jia Huang, Min-min Huang","doi":"10.1145/3583788.3583789","DOIUrl":"https://doi.org/10.1145/3583788.3583789","url":null,"abstract":"The factor score method is a commonly used method to solve the problem of location, but the implementation process is greatly influenced by the subjective evaluation of experts, and the calculation results lack objectivity and accuracy. In order to overcome the above-mentioned shortcomings, this paper proposes a fuzzy multi-criteria decision-making model based on linguistic distribution evaluation method, entropy weight method and TODIM (interactive multi-criteria decision-making) method for location research. Firstly, this paper uses the linguistic distribution evaluation method to obtain the expert evaluation information, then uses the entropy weight method to calculate the weight of each influencing factor, and finally uses the TODIM method to determine the optimal location.Through case comparison and analysis, it is found that the model can effectively reduce the uncertainty of expert evaluation and improve the reliability of location results.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"172 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134356760","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}
Quantitative Structure-Activity Relationships (QSAR), which aims to estimate the estrogen receptor alpha (ERα) activity of compounds through their chemical features and ERα, is a fundamental part in the process of drug discovery for breast cancer treatment. Due to the variety of data properties, the building of a suitable QSAR model is a challenging task. Meanwhile, the challenge of QSAR lies in the complexity of compound molecular descriptors which make it difficult to screen robust molecular descriptors. Previous studies select molecular descriptors manually based on expert knowledge and experience. However, they are highly subjective which could lead to ineffectiveness of molecular descriptors. In this paper, a novel approach is presented to address the problems in the context of regression modelling and feature selection. Firstly, two filtered and two embedded scoring metrics are proposed to jointly sort and select the most relevant and robust molecular descriptors. Then the selected features are used to build the supervised data-driven model, namely eXtreme Gradient Boosting (XGBoost) algorithm. Experimental results show that our selected molecular descriptors can give good predictions to the target ERα bioactivity and our regression approach outperform formal models.
{"title":"Quantitative Structure-Activity Relationships of Estrogen Receptor Alpha Based on Molecular Descriptors Selection and Extreme Gradient Boosting","authors":"Shaotong Liu, Zhewei Xu, Dongsheng Ye","doi":"10.1145/3583788.3583807","DOIUrl":"https://doi.org/10.1145/3583788.3583807","url":null,"abstract":"Quantitative Structure-Activity Relationships (QSAR), which aims to estimate the estrogen receptor alpha (ERα) activity of compounds through their chemical features and ERα, is a fundamental part in the process of drug discovery for breast cancer treatment. Due to the variety of data properties, the building of a suitable QSAR model is a challenging task. Meanwhile, the challenge of QSAR lies in the complexity of compound molecular descriptors which make it difficult to screen robust molecular descriptors. Previous studies select molecular descriptors manually based on expert knowledge and experience. However, they are highly subjective which could lead to ineffectiveness of molecular descriptors. In this paper, a novel approach is presented to address the problems in the context of regression modelling and feature selection. Firstly, two filtered and two embedded scoring metrics are proposed to jointly sort and select the most relevant and robust molecular descriptors. Then the selected features are used to build the supervised data-driven model, namely eXtreme Gradient Boosting (XGBoost) algorithm. Experimental results show that our selected molecular descriptors can give good predictions to the target ERα bioactivity and our regression approach outperform formal models.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121095932","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}
Diabetes is a common disease, and due to the increasing incidence year by year. But most diabetics can not be easily detected in the early stage, since the symptoms are not obvious. The objective of this study is to propose a machine-learning method based on unsupervised clustering to improve the accuracy of diabetes detection. Due to massive unlabeled data sets and the problems in the traditional K-means clustering algorithms, we adopt the Fuzzy c-means clustering algorithm with an improvement on the calculation of parameter m. Our method includes a combination of the principal component analysis(PCA), an improved Fuzzy c-means (FCM) clustering algorithm, and K-nearest neighbor(KNN) classification algorithm optimized with K value. After 10 times 10-fold cross-validation, the average accuracy of the proposed method reaches 99.31%, which is higher than that of other machine learning models. Therefore, our method is proven to be more suitable for detecting diabetes. At the same time, further experiments on a new data set validate the applicability of our method in a more practical way for the diabetes detection.
{"title":"A Hybrid Machine Learning Method for Diabetes Detection based on Unsupervised Clustering","authors":"Junhong Liu, Bo Peng, Zezhao Yin","doi":"10.1145/3583788.3583809","DOIUrl":"https://doi.org/10.1145/3583788.3583809","url":null,"abstract":"Diabetes is a common disease, and due to the increasing incidence year by year. But most diabetics can not be easily detected in the early stage, since the symptoms are not obvious. The objective of this study is to propose a machine-learning method based on unsupervised clustering to improve the accuracy of diabetes detection. Due to massive unlabeled data sets and the problems in the traditional K-means clustering algorithms, we adopt the Fuzzy c-means clustering algorithm with an improvement on the calculation of parameter m. Our method includes a combination of the principal component analysis(PCA), an improved Fuzzy c-means (FCM) clustering algorithm, and K-nearest neighbor(KNN) classification algorithm optimized with K value. After 10 times 10-fold cross-validation, the average accuracy of the proposed method reaches 99.31%, which is higher than that of other machine learning models. Therefore, our method is proven to be more suitable for detecting diabetes. At the same time, further experiments on a new data set validate the applicability of our method in a more practical way for the diabetes detection.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125685803","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}
Nowadays, violence in movies and in society is on the rise, which has a significant impact on children, particularly adolescents. The prevalence of school violence is increasing and it is becoming a concern for schools, families, and society as a whole. However, because the school violence detection system has not yet been developed, our lab created VSiSGU data based on the collection of camera data from within the school as well as data from social networks. There are also many techniques for processing continuous image sequence data from cameras in order to detect school violence. As a result, we propose a method for improving performance by selecting frames at the l, l+k, l+2k,..., l+nk positions in the videos to train. After that, we use the VGGNet algorithm combined with RNN to develop a training model on the above data. The evaluation results show that our proposed method is more efficient in terms of time and still ensures higher or equivalent accuracy than the traditional sampling method.
{"title":"Proposal to Improve The Classification of School Violence","authors":"Ha Duong Ngo, Y. Tran","doi":"10.1145/3583788.3583811","DOIUrl":"https://doi.org/10.1145/3583788.3583811","url":null,"abstract":"Nowadays, violence in movies and in society is on the rise, which has a significant impact on children, particularly adolescents. The prevalence of school violence is increasing and it is becoming a concern for schools, families, and society as a whole. However, because the school violence detection system has not yet been developed, our lab created VSiSGU data based on the collection of camera data from within the school as well as data from social networks. There are also many techniques for processing continuous image sequence data from cameras in order to detect school violence. As a result, we propose a method for improving performance by selecting frames at the l, l+k, l+2k,..., l+nk positions in the videos to train. After that, we use the VGGNet algorithm combined with RNN to develop a training model on the above data. The evaluation results show that our proposed method is more efficient in terms of time and still ensures higher or equivalent accuracy than the traditional sampling method.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128089421","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}
Although depth completion has achieved remarkable performance relying on deep learning in recent years, these models tend to suffer a performance degradation when exposed to new environments. Online adaptation, where the model is trained in a self-supervised manner during testing, seems a promising technique to alleviate the drop. However, continuous online adaptation may cause the model to over-adapt and miss the optimal parameters, resulting in oscillation or even degradation of the model performance, in addition to wasting computational resources. Therefore, this paper proposes an adaptive online adaptation framework to make model adaptively trigger online adaptation when encountering novel environments and stop adaptation when model has adapted to the current environment. In detail, we design a trigger to detect the familiarity of model to the current scenario based on image similarity and then launch online adaptation when the scenario is novel. Besides, we elaborate a stopper to monitor the error between prediction and depth input and convert online adaptation to inference when online adaptation does not bring improvement for model. Experimental results demonstrate that our method improves the accuracy of model prediction and increases average running speed of the model on each frame in online adaptation.
{"title":"Self-supervised Depth Completion with Adaptive Online Adaptation","authors":"Yang Chen, Yang Tan","doi":"10.1145/3583788.3583813","DOIUrl":"https://doi.org/10.1145/3583788.3583813","url":null,"abstract":"Although depth completion has achieved remarkable performance relying on deep learning in recent years, these models tend to suffer a performance degradation when exposed to new environments. Online adaptation, where the model is trained in a self-supervised manner during testing, seems a promising technique to alleviate the drop. However, continuous online adaptation may cause the model to over-adapt and miss the optimal parameters, resulting in oscillation or even degradation of the model performance, in addition to wasting computational resources. Therefore, this paper proposes an adaptive online adaptation framework to make model adaptively trigger online adaptation when encountering novel environments and stop adaptation when model has adapted to the current environment. In detail, we design a trigger to detect the familiarity of model to the current scenario based on image similarity and then launch online adaptation when the scenario is novel. Besides, we elaborate a stopper to monitor the error between prediction and depth input and convert online adaptation to inference when online adaptation does not bring improvement for model. Experimental results demonstrate that our method improves the accuracy of model prediction and increases average running speed of the model on each frame in online adaptation.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130890040","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 problems of low accuracy and slow recognition efficiency of the traditional traffic sign recognition algorithm in complex environment, a deep learning traffic sign recognition method based on YOLOv5 is proposed. Firstly, the Chinese traffic sign data set TT100K is randomly divided into training set and test set. Convolutional neural network YOLOv4 and convolutional neural network YOLOv5 are used to train respectively on the training set, so as to build the prediction model of traffic signs. Then the trained model is validated on the test set. Through the evaluation of the experimental, it is found that compared with YOLOv4 model, YOLOv5 model has higher recognition accuracy and faster recognition speed.
{"title":"A Traffic Sign Recognition Method Based on YOLOv5 Deep Learning Algorithm","authors":"Yinqing Tang, Benguo Yu, Anran Wang, Fengning Liu","doi":"10.1145/3583788.3583810","DOIUrl":"https://doi.org/10.1145/3583788.3583810","url":null,"abstract":"Aiming at the problems of low accuracy and slow recognition efficiency of the traditional traffic sign recognition algorithm in complex environment, a deep learning traffic sign recognition method based on YOLOv5 is proposed. Firstly, the Chinese traffic sign data set TT100K is randomly divided into training set and test set. Convolutional neural network YOLOv4 and convolutional neural network YOLOv5 are used to train respectively on the training set, so as to build the prediction model of traffic signs. Then the trained model is validated on the test set. Through the evaluation of the experimental, it is found that compared with YOLOv4 model, YOLOv5 model has higher recognition accuracy and faster recognition speed.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132921890","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}
Jianhui Wang, Biao Jie, Xingyu Zhang, Wen J. Li, Zhaoxiang Wu, Yang Yang
Dynamic functional connectivity network (DFCN) derived from resting-state functional magnetic resonance imaging (rs-fMRI), which characterizes the dynamic interaction between brain regions, has been applied to classification of brain diseases. However, existing studies usually focus on dynamic changes of low-order (i.e., pairwise) correlation of brain regions, thus neglecting their high-order dynamic information that could be important for brain disease diagnosis. Therefore, in this paper, we first propose a novel sparse learning based high-order DFCNs construction method, and then build a novel learning framework to extract high-level and high-order temporal features from the constructed high-order DFCNs for brain disease classification. The experimental results on 174 subjects from from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) demonstrate the effectiveness of our proposed method in comparison with state-of-the-art methods.
{"title":"Sparse-learning-based High-order Dynamic Functional Connectivity Networks for Brain Disease Classification","authors":"Jianhui Wang, Biao Jie, Xingyu Zhang, Wen J. Li, Zhaoxiang Wu, Yang Yang","doi":"10.1145/3583788.3583812","DOIUrl":"https://doi.org/10.1145/3583788.3583812","url":null,"abstract":"Dynamic functional connectivity network (DFCN) derived from resting-state functional magnetic resonance imaging (rs-fMRI), which characterizes the dynamic interaction between brain regions, has been applied to classification of brain diseases. However, existing studies usually focus on dynamic changes of low-order (i.e., pairwise) correlation of brain regions, thus neglecting their high-order dynamic information that could be important for brain disease diagnosis. Therefore, in this paper, we first propose a novel sparse learning based high-order DFCNs construction method, and then build a novel learning framework to extract high-level and high-order temporal features from the constructed high-order DFCNs for brain disease classification. The experimental results on 174 subjects from from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) demonstrate the effectiveness of our proposed method in comparison with state-of-the-art methods.","PeriodicalId":292167,"journal":{"name":"Proceedings of the 2023 7th International Conference on Machine Learning and Soft Computing","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131741016","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：