Pub Date : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00042
Yinyihong Liu
Being one of the largest online accommodation booking platforms, Airbnb has many hosts who are seeking for more proper prices to increase their booking rate. To develop a good pricing prediction model, this paper has employed machine learning models including KNN, MLR, LASSO regression, Ridge regression, Random Forest, Gradient Boosting and XGBoost etc. While past studies on Airbnb pricing have applied quantitative pricing, some face the problems that the models are not robust enough and some face the problem of not training the model plentily. To fill this gap, we give careful consideration in exploratory data analysis to make the dataset more reasonable, apply many robust models ranging from regularized regression to ensemble models and use cross validation and random search to tune each parameter in each model. In this way, we not only select XGBoost as the best model for price prediction with R2 score 0.6321, but also uncover the features which have statistical significance with the target price.
{"title":"Airbnb Pricing Based on Statistical Machine Learning Models","authors":"Yinyihong Liu","doi":"10.1109/CONF-SPML54095.2021.00042","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00042","url":null,"abstract":"Being one of the largest online accommodation booking platforms, Airbnb has many hosts who are seeking for more proper prices to increase their booking rate. To develop a good pricing prediction model, this paper has employed machine learning models including KNN, MLR, LASSO regression, Ridge regression, Random Forest, Gradient Boosting and XGBoost etc. While past studies on Airbnb pricing have applied quantitative pricing, some face the problems that the models are not robust enough and some face the problem of not training the model plentily. To fill this gap, we give careful consideration in exploratory data analysis to make the dataset more reasonable, apply many robust models ranging from regularized regression to ensemble models and use cross validation and random search to tune each parameter in each model. In this way, we not only select XGBoost as the best model for price prediction with R2 score 0.6321, but also uncover the features which have statistical significance with the target price.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125235590","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00046
Zhi Yi, Yuyang Wang
For the treatment of Interstitial Lung Disease, it is crucial to have an early diagnosis. However, doctors still have a lot of controversy in the diagnosis of lung nodules even with today’s highly developed medical imaging technology. In this article, we summarized the five major challenges we face in medical image recognition and systematically listed the applications from traditional image recognition technology to deep learning in lung CT image recognition. Compared to the traditional convolutional neural network built and trained from scratch, it is beneficial to apply transfer learning to the recognition of lung nodules. Transfer learning focus on transferring knowledge from previous well-trained task to target learning task. Transferring means pretrained networks utilize fine-tuning to reduce iteration times of weight so that it can cope with the problem of lack of high quality images. Various experiments demonstrate that transfer learning performances better than traditional convolutional neural network under complicated circumstances of image recognition such as medical images. In this article, transfer learning is classified into 3 types: inductive transfer learning, transductive transfer learning and unsupervised transfer learning. The main difference between them is label quantity of target training set. Inductive transfer learning highly depends on feature engineering. Compared to it, training sets of two remaining has few labels. However, transductive transfer learning and unsupervised transfer learning are unstable while facing sophisticated cases.
{"title":"Transfer Learning on Interstitial Lung Disease Classification","authors":"Zhi Yi, Yuyang Wang","doi":"10.1109/CONF-SPML54095.2021.00046","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00046","url":null,"abstract":"For the treatment of Interstitial Lung Disease, it is crucial to have an early diagnosis. However, doctors still have a lot of controversy in the diagnosis of lung nodules even with today’s highly developed medical imaging technology. In this article, we summarized the five major challenges we face in medical image recognition and systematically listed the applications from traditional image recognition technology to deep learning in lung CT image recognition. Compared to the traditional convolutional neural network built and trained from scratch, it is beneficial to apply transfer learning to the recognition of lung nodules. Transfer learning focus on transferring knowledge from previous well-trained task to target learning task. Transferring means pretrained networks utilize fine-tuning to reduce iteration times of weight so that it can cope with the problem of lack of high quality images. Various experiments demonstrate that transfer learning performances better than traditional convolutional neural network under complicated circumstances of image recognition such as medical images. In this article, transfer learning is classified into 3 types: inductive transfer learning, transductive transfer learning and unsupervised transfer learning. The main difference between them is label quantity of target training set. Inductive transfer learning highly depends on feature engineering. Compared to it, training sets of two remaining has few labels. However, transductive transfer learning and unsupervised transfer learning are unstable while facing sophisticated cases.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130160801","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00010
Chen Xiong
In this paper, based on LD3320 non-specific person speech recognition chip, Arduino UNO R3 MCU, LX-225 serial bus intelligent steering gear, an intelligent trash can is designed, which realizes the functions of voice input of garbage name, intelligent retrieval of garbage type, intelligent opening and closing of garbage can cover, etc.
本文基于LD3320非特定人物语音识别芯片、Arduino UNO R3单片机、LX-225串行总线智能舵机，设计了一种智能垃圾桶，实现了垃圾名称语音输入、垃圾类型智能检索、垃圾桶盖智能开合等功能。
{"title":"Design of Intelligent Garbage Classification Bin Based on LD3320","authors":"Chen Xiong","doi":"10.1109/CONF-SPML54095.2021.00010","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00010","url":null,"abstract":"In this paper, based on LD3320 non-specific person speech recognition chip, Arduino UNO R3 MCU, LX-225 serial bus intelligent steering gear, an intelligent trash can is designed, which realizes the functions of voice input of garbage name, intelligent retrieval of garbage type, intelligent opening and closing of garbage can cover, etc.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125437652","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00060
Kai Wang
Deep Learning (DL) has been proven to be a promising technique for image analysis tasks such as image classification and object recognition. Compared with other fields, the accuracy of DL tasks in medical imaging depends heavily on the dataset volume. However, DL has been suffering from the problem of small sample datasets caused by a variety of ethical and financial reasons in medical imaging. Data augmentation and transfer learning are the two most commonly used approaches to enhance the practicability of the DL algorithms in medical imaging. This article discusses the data augmentation methods including image manipulation and generative adversarial networks. Feature-extracting and fine-tuning methods of transfer learning are also discussed. Finally, the paper mentions the real-life applications of many architectures, advantages and disadvantages, and future works.
{"title":"An Overview of Deep Learning Based Small Sample Medical Imaging Classification","authors":"Kai Wang","doi":"10.1109/CONF-SPML54095.2021.00060","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00060","url":null,"abstract":"Deep Learning (DL) has been proven to be a promising technique for image analysis tasks such as image classification and object recognition. Compared with other fields, the accuracy of DL tasks in medical imaging depends heavily on the dataset volume. However, DL has been suffering from the problem of small sample datasets caused by a variety of ethical and financial reasons in medical imaging. Data augmentation and transfer learning are the two most commonly used approaches to enhance the practicability of the DL algorithms in medical imaging. This article discusses the data augmentation methods including image manipulation and generative adversarial networks. Feature-extracting and fine-tuning methods of transfer learning are also discussed. Finally, the paper mentions the real-life applications of many architectures, advantages and disadvantages, and future works.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126077456","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00064
J. Vugrin, S. Lončarić
Face analysis is a broad and well-established research area whose main focus is put on face detection, segmentation, recognition and facial features extraction. A crucial prerequisite to face analysis algorithms properly work is to have an input image of high quality with similar properties in different conditions. For this reason, near-infrared images are used due to being more robust to change in lighting conditions and time of day than the visible light spectrum images. Automatic brightness control is used to properly adjust scene brightness to extract useful information. A novel algorithm implementation for the automatic brightness control is proposed based on a split range feedback controller with a camera occlusion detection included. The proposed algorithm is accurate, fast and suitable for real-time embedded system implementation.
{"title":"Automatic Brightness Control for Face Analysis in Near-Infrared Spectrum","authors":"J. Vugrin, S. Lončarić","doi":"10.1109/CONF-SPML54095.2021.00064","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00064","url":null,"abstract":"Face analysis is a broad and well-established research area whose main focus is put on face detection, segmentation, recognition and facial features extraction. A crucial prerequisite to face analysis algorithms properly work is to have an input image of high quality with similar properties in different conditions. For this reason, near-infrared images are used due to being more robust to change in lighting conditions and time of day than the visible light spectrum images. Automatic brightness control is used to properly adjust scene brightness to extract useful information. A novel algorithm implementation for the automatic brightness control is proposed based on a split range feedback controller with a camera occlusion detection included. The proposed algorithm is accurate, fast and suitable for real-time embedded system implementation.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"125 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131593607","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00030
Lilan Wen
With the increasing demand for computing speed and real-time data processing in various fields, deep learning and convolutional neural networks are more and more widely used in the field of computer vision. FPGA-based deep convolutional neural networks (CNN) have been proposed and developed rapidly due to its high parallel processing ability, portability, and low power consumption. To further improve the network efficiency, this paper studies the software acceleration tool Vivado HLS provided by Xilinx, the quantification and pruning of convolution neural network model, which can effectively optimize the network model and accelerate the reasoning process.
{"title":"FPGA-Based Deep Convolutional Neural Network Optimization Method","authors":"Lilan Wen","doi":"10.1109/CONF-SPML54095.2021.00030","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00030","url":null,"abstract":"With the increasing demand for computing speed and real-time data processing in various fields, deep learning and convolutional neural networks are more and more widely used in the field of computer vision. FPGA-based deep convolutional neural networks (CNN) have been proposed and developed rapidly due to its high parallel processing ability, portability, and low power consumption. To further improve the network efficiency, this paper studies the software acceleration tool Vivado HLS provided by Xilinx, the quantification and pruning of convolution neural network model, which can effectively optimize the network model and accelerate the reasoning process.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134443768","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00009
Longshuai Sheng, Ce Li, Yihan Tian
Weakly supervised action segmentation has been extensively studied to get the category and start time of actions that occur in videos, but it remains an unsolved issue because of lacking great annotation data in video analysis. To handle this issue, weakly supervised action segmentation only uses the action annotation on the whole sequence in a long video instead of specific labeling of each frame, which greatly reduces the difficulty of obtaining video datasets. However, the task remains challenging for the complex temporal length partition of actions in the videos. In this paper, we make use of the Viterbi algorithm to generate an initial action segmentation as the baseline and then design a new coarse-to-fine loss function to refine the length partition and learn the scores of valid and invalid segmentation routes respectively. The new coarse-to-fine loss is learned in the pipeline to reduce the weight of invalid segmentation routes and obtain the best video segmentation. Comparing with the state-of-the-art (SOTA) methods, the experiments on the breakfast and 50 salads datasets show that our fine partition model and coarse-to-fine loss function can be used to obtain higher frame accuracy and significantly reduce the time spent for action segmentation.
{"title":"Coarse-to-Fine Loss Based On Viterbi Algorithm for Weakly Supervised Action Segmentation","authors":"Longshuai Sheng, Ce Li, Yihan Tian","doi":"10.1109/CONF-SPML54095.2021.00009","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00009","url":null,"abstract":"Weakly supervised action segmentation has been extensively studied to get the category and start time of actions that occur in videos, but it remains an unsolved issue because of lacking great annotation data in video analysis. To handle this issue, weakly supervised action segmentation only uses the action annotation on the whole sequence in a long video instead of specific labeling of each frame, which greatly reduces the difficulty of obtaining video datasets. However, the task remains challenging for the complex temporal length partition of actions in the videos. In this paper, we make use of the Viterbi algorithm to generate an initial action segmentation as the baseline and then design a new coarse-to-fine loss function to refine the length partition and learn the scores of valid and invalid segmentation routes respectively. The new coarse-to-fine loss is learned in the pipeline to reduce the weight of invalid segmentation routes and obtain the best video segmentation. Comparing with the state-of-the-art (SOTA) methods, the experiments on the breakfast and 50 salads datasets show that our fine partition model and coarse-to-fine loss function can be used to obtain higher frame accuracy and significantly reduce the time spent for action segmentation.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133607386","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00028
L. Liu
In recent years, UAVs have received more and more attention from countries all over the world. The diverse functions and unique advantages have caused the world to set off an upsurge in UAV development. Taking UAV as the research object, this thesis mainly studies the design of the flight control law of the UAV flight control system based on PID control, so as to facilitate the flight simulation of UAV. In response to this problem, the longitudinal flight control law and the lateral flight control law are designed to maintain and control the movement of the UAV’s altitude, pitching angle, roll angle and course angle.
{"title":"Design of UAV Flight Control Law Based on PID Control","authors":"L. Liu","doi":"10.1109/CONF-SPML54095.2021.00028","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00028","url":null,"abstract":"In recent years, UAVs have received more and more attention from countries all over the world. The diverse functions and unique advantages have caused the world to set off an upsurge in UAV development. Taking UAV as the research object, this thesis mainly studies the design of the flight control law of the UAV flight control system based on PID control, so as to facilitate the flight simulation of UAV. In response to this problem, the longitudinal flight control law and the lateral flight control law are designed to maintain and control the movement of the UAV’s altitude, pitching angle, roll angle and course angle.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131453126","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00071
Ning Qin, Xinyu Zhou, Jiaqi Wang, Chujie Shen
Bayesian optimization(BO) is a global optimization problem. It is an important approach in machine learning, hyperparameter tuning and other fields such as drug discovery. BO consists of two main parts which are probabilistic model for the objective function and acquisition function. This paper mainly focused on assessing the strengths and weaknesses of two different probabilistic models which are Gaussian Process (GP) and Random Forests (RF). This paper illustrated several results, which indicated the performance of each probabilistic model and helped us find the optimal model corresponding to each benchmark function. RF will be preferred if the function is smooth. GP will be preferred if the function has many local minima. Moreover, implementability of other probabilistic models were discussed in this paper.
{"title":"Bayesian Optimization: Model Comparison With Different Benchmark Functions","authors":"Ning Qin, Xinyu Zhou, Jiaqi Wang, Chujie Shen","doi":"10.1109/CONF-SPML54095.2021.00071","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00071","url":null,"abstract":"Bayesian optimization(BO) is a global optimization problem. It is an important approach in machine learning, hyperparameter tuning and other fields such as drug discovery. BO consists of two main parts which are probabilistic model for the objective function and acquisition function. This paper mainly focused on assessing the strengths and weaknesses of two different probabilistic models which are Gaussian Process (GP) and Random Forests (RF). This paper illustrated several results, which indicated the performance of each probabilistic model and helped us find the optimal model corresponding to each benchmark function. RF will be preferred if the function is smooth. GP will be preferred if the function has many local minima. Moreover, implementability of other probabilistic models were discussed in this paper.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132820491","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 : 2021-11-01DOI: 10.1109/CONF-SPML54095.2021.00053
Xin Li, Ce Li, Xianlong Wei, Feng Yang
As the key application in video analysis for human computer interaction (HCI), the problem of skeleton-based action recognition has been solved by some researchers with graph neural networks, but it remains an unsolved issue on complex variations of spatiotemporal dependence across skeleton joints flow. A newly dynamic spatio-temporal graph structure learning method, manifold guided graph neural networks (MGNN), was proposed to solve this problem. In MGNN, a novel manifold guided graph updating mechanism is built based on the baseline graph neural network to further describe the spatio-temporal dependence. With the manifold guided multi-scale skeleton graph, the proposed MGNN is further trained with two streams of joint and bone to improve the efficiency, which forms a single network seamlessly and enables it be trained in a same umbrella. Comparing with the existing methods, MGNN has been proved that it yields better performance on challenging datasets: NTU RGB+D 60 and Kinetics 400.
{"title":"Manifold Guided Graph Neural Networks for Skeleton-based Action Recognition in Human Computer Interaction Videos","authors":"Xin Li, Ce Li, Xianlong Wei, Feng Yang","doi":"10.1109/CONF-SPML54095.2021.00053","DOIUrl":"https://doi.org/10.1109/CONF-SPML54095.2021.00053","url":null,"abstract":"As the key application in video analysis for human computer interaction (HCI), the problem of skeleton-based action recognition has been solved by some researchers with graph neural networks, but it remains an unsolved issue on complex variations of spatiotemporal dependence across skeleton joints flow. A newly dynamic spatio-temporal graph structure learning method, manifold guided graph neural networks (MGNN), was proposed to solve this problem. In MGNN, a novel manifold guided graph updating mechanism is built based on the baseline graph neural network to further describe the spatio-temporal dependence. With the manifold guided multi-scale skeleton graph, the proposed MGNN is further trained with two streams of joint and bone to improve the efficiency, which forms a single network seamlessly and enables it be trained in a same umbrella. Comparing with the existing methods, MGNN has been proved that it yields better performance on challenging datasets: NTU RGB+D 60 and Kinetics 400.","PeriodicalId":415094,"journal":{"name":"2021 International Conference on Signal Processing and Machine Learning (CONF-SPML)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115853720","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：