Pub Date : 2023-09-20DOI: 10.20965/jaciii.2023.p0967
Yi Liu, Jianzhang Li, Dewen Cui, Eri Sato-Shimokawara
Affective computing focuses on recognizing emotions using a combination of psychology, computer science, and biomedical engineering. With virtual reality (VR) becoming more widely accessible, affective computing has become increasingly important for supporting social interactions on online virtual platforms. However, accurately estimating a person’s emotional state in VR is challenging because it differs from real-world conditions, such as the unavailability of facial expressions. This research proposes a self-training method that uses unlabeled data and a reinforcement learning approach to select and label data more accurately. Experiments on a dataset of dialogues of VR players show that the proposed method achieved an accuracy of over 80% on dominance and arousal labels and outperformed previous techniques in the few-shot classification of emotions based on physiological signals.
{"title":"Multi-Modal Emotion Classification in Virtual Reality Using Reinforced Self-Training","authors":"Yi Liu, Jianzhang Li, Dewen Cui, Eri Sato-Shimokawara","doi":"10.20965/jaciii.2023.p0967","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0967","url":null,"abstract":"Affective computing focuses on recognizing emotions using a combination of psychology, computer science, and biomedical engineering. With virtual reality (VR) becoming more widely accessible, affective computing has become increasingly important for supporting social interactions on online virtual platforms. However, accurately estimating a person’s emotional state in VR is challenging because it differs from real-world conditions, such as the unavailability of facial expressions. This research proposes a self-training method that uses unlabeled data and a reinforcement learning approach to select and label data more accurately. Experiments on a dataset of dialogues of VR players show that the proposed method achieved an accuracy of over 80% on dominance and arousal labels and outperformed previous techniques in the few-shot classification of emotions based on physiological signals.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264986","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0790
Qiwei Yu, Yaping Dai, Kaoru Hirota, Shuai Shao, Wei Dai
A shuffle graph convolutional network (Shuffle-GCN) is proposed to recognize human action by analyzing skeleton data. It uses channel split and channel shuffle operations to process multi-feature channels of skeleton data, which reduces the computational cost of graph convolution operation. Compared with the classical two-stream adaptive graph convolutional network model, the proposed method achieves a higher precision with 1/3 of the floating-point operations (FLOPs). Even more, a channel-level topology modeling method is designed to extract more motion information of human skeleton by learning the graph topology from different channels dynamically. The performance of Shuffle-GCN is tested under 56,880 action clips from the NTU RGB+D dataset with the accuracy 96.0% and the computational complexity 12.8 GFLOPs. The proposed method offers feasible solutions for developing practical applications of action recognition.
{"title":"Shuffle Graph Convolutional Network for Skeleton-Based Action Recognition","authors":"Qiwei Yu, Yaping Dai, Kaoru Hirota, Shuai Shao, Wei Dai","doi":"10.20965/jaciii.2023.p0790","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0790","url":null,"abstract":"A shuffle graph convolutional network (Shuffle-GCN) is proposed to recognize human action by analyzing skeleton data. It uses channel split and channel shuffle operations to process multi-feature channels of skeleton data, which reduces the computational cost of graph convolution operation. Compared with the classical two-stream adaptive graph convolutional network model, the proposed method achieves a higher precision with 1/3 of the floating-point operations (FLOPs). Even more, a channel-level topology modeling method is designed to extract more motion information of human skeleton by learning the graph topology from different channels dynamically. The performance of Shuffle-GCN is tested under 56,880 action clips from the NTU RGB+D dataset with the accuracy 96.0% and the computational complexity 12.8 GFLOPs. The proposed method offers feasible solutions for developing practical applications of action recognition.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264995","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0801
Jing Hu, Changjuan Lao, Xiaomeng Su
This paper focuses on latecomers in technology standards alliances, a topic that has received limited attention in academia. Although latecomers have disadvantages in terms of technology and market access, they possess a strong desire to catch up technologically. As a fundamental characteristic of an alliance, alliance routines serve as a source of innovation for members and a basic unit of analysis, providing a new perspective for understanding and researching the technological catch-up of latecomers. In this paper, a questionnaire survey is conducted among 83 latecomer enterprises in a technology standards alliance. The survey covers strategic emerging industries, such as new energy, new-generation information technology, new materials, and high-end manufacturing. Then, hierarchical regression is performed for hypothesis testing. The research shows that the three dimensions of alliance routines significantly promote the utilization of technological catch-up by latecomer enterprises. However, the impact on the exploration technological catch-up varies, where the action logic promotes exploration technological catch-up while implicit norms hinder exploratory innovation with an inverted U-shaped impact. The absorptive capacity strengthens the positive impact of the three alliance routines on the exploitative innovation of latecomer enterprises and strengthens the inverted U-shaped relationship between mutual consensus and their exploratory innovation. Nonetheless, it fails to play a significant regulatory role in action logic or in the relationship between implicit norms and exploratory innovation of latecomers.
{"title":"Research on the Influence of Alliance Routines on the Ambidextrous Technological Catch-Up of Latecomers to Technology Standards Alliances","authors":"Jing Hu, Changjuan Lao, Xiaomeng Su","doi":"10.20965/jaciii.2023.p0801","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0801","url":null,"abstract":"This paper focuses on latecomers in technology standards alliances, a topic that has received limited attention in academia. Although latecomers have disadvantages in terms of technology and market access, they possess a strong desire to catch up technologically. As a fundamental characteristic of an alliance, alliance routines serve as a source of innovation for members and a basic unit of analysis, providing a new perspective for understanding and researching the technological catch-up of latecomers. In this paper, a questionnaire survey is conducted among 83 latecomer enterprises in a technology standards alliance. The survey covers strategic emerging industries, such as new energy, new-generation information technology, new materials, and high-end manufacturing. Then, hierarchical regression is performed for hypothesis testing. The research shows that the three dimensions of alliance routines significantly promote the utilization of technological catch-up by latecomer enterprises. However, the impact on the exploration technological catch-up varies, where the action logic promotes exploration technological catch-up while implicit norms hinder exploratory innovation with an inverted U-shaped impact. The absorptive capacity strengthens the positive impact of the three alliance routines on the exploitative innovation of latecomer enterprises and strengthens the inverted U-shaped relationship between mutual consensus and their exploratory innovation. Nonetheless, it fails to play a significant regulatory role in action logic or in the relationship between implicit norms and exploratory innovation of latecomers.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136265000","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}
Information and communication technology-based monitoring services for the elderly can be improved by using a monitoring function that automatically detects cognitive decline. Therefore, we developed a system that provides a brief screening for mild cognitive impairment (MCI). This screening is performed by measuring the performance of two tasks completed simultaneously on a tablet device, such as drawing a spiral and counting color changes. First, we used this system to measure and analyze three groups: elderly persons suspected of having MCI, elderly persons with no cognitive impairment, and healthy young adults. We extracted evaluation parameters relevant for determining MCI through analysis. Furthermore, we developed a detection method using the evaluation parameters and performed an evaluation. Cross-validation revealed that the sensitivity and specificity of this method were 90.0% and 75.0%, respectively.
{"title":"Detection of Mild Cognitive Impairment Using a Spiral Drawing Test","authors":"Katsuya Fujiwara, Kenta Matsuhashi, Kazutaka Mitobe","doi":"10.20965/jaciii.2023.p0907","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0907","url":null,"abstract":"Information and communication technology-based monitoring services for the elderly can be improved by using a monitoring function that automatically detects cognitive decline. Therefore, we developed a system that provides a brief screening for mild cognitive impairment (MCI). This screening is performed by measuring the performance of two tasks completed simultaneously on a tablet device, such as drawing a spiral and counting color changes. First, we used this system to measure and analyze three groups: elderly persons suspected of having MCI, elderly persons with no cognitive impairment, and healthy young adults. We extracted evaluation parameters relevant for determining MCI through analysis. Furthermore, we developed a detection method using the evaluation parameters and performed an evaluation. Cross-validation revealed that the sensitivity and specificity of this method were 90.0% and 75.0%, respectively.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"165 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264237","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0886
Shinji Uchinoura, Junichi Miyao, Takio Kurita
This paper proposes a two-step detector called segmented object detection, whose performance is improved by masking the background region. Previous single-stage object detection methods suffer from the problem of imbalance between foreground and background classes, where the background occupies more regions in the image than the foreground. Thus, the loss from the background is firmly incorporated into the training. RetinaNet addresses this problem with Focal Loss, which focuses on foreground loss. Therefore, we propose a method that generates probability maps using instance segmentation in the first step and feeds back the generated maps as background masks in the second step as prior knowledge to reduce the influence of the background and enhance foreground training. We confirm that the detector can improve the accuracy by adding instance segmentation information to both the input and output rather than only to the output results. On the Cityscapes dataset, our method outperforms the state-of-the-art methods.
{"title":"An Object Detection Method Using Probability Maps for Instance Segmentation to Mask Background","authors":"Shinji Uchinoura, Junichi Miyao, Takio Kurita","doi":"10.20965/jaciii.2023.p0886","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0886","url":null,"abstract":"This paper proposes a two-step detector called segmented object detection, whose performance is improved by masking the background region. Previous single-stage object detection methods suffer from the problem of imbalance between foreground and background classes, where the background occupies more regions in the image than the foreground. Thus, the loss from the background is firmly incorporated into the training. RetinaNet addresses this problem with Focal Loss, which focuses on foreground loss. Therefore, we propose a method that generates probability maps using instance segmentation in the first step and feeds back the generated maps as background masks in the second step as prior knowledge to reduce the influence of the background and enhance foreground training. We confirm that the detector can improve the accuracy by adding instance segmentation information to both the input and output rather than only to the output results. On the Cityscapes dataset, our method outperforms the state-of-the-art methods.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264247","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0848
Naomi Yagi, Yutaka Hata, Yoshitada Sakai
Articulation disorder is a condition in which the mouth, tongue, vocal cords, and other parts of the body that play an important role in producing voice are damaged, resulting in the inability to produce sound. To diagnose articulation disorders, the movement and shape of each organ concerned with pronunciation are examined. If necessary, the underlying disease or disorder should be managed properly. In it, a speech therapist tests your pronunciation. The observation of conversation and the examination of the pronunciation of each syllable are used to distinguish between mistakes and the degree of articulation disorder. However, these processes are time consuming and labor intensive and are subjective judgments by experts. Therefore, it is important to investigate the characteristics of vocal signals by acoustic analysis of speech objectively. In this study, we focused on fluctuations in the period and amplitude of speech signals and predicted a model for detecting abnormal articulations using fluctuation measurement of the voice data in six healthy subjects and nine patients with an articulation disorder. We used inverse probability of treatment weighting to match the variability for the two groups using the inverse of propensity scores. As the results, the classification performance area under the curve was 0.781 (sensitivity: 0.781, specificity: 0.680) for healthy subjects and patients. We conclude that acoustic analyzing techniques are useful for diagnosing and treating articulation disorders.
{"title":"Abnormal Articulation Detecting Model with Fluctuation Measurements Using Acoustic Analysis","authors":"Naomi Yagi, Yutaka Hata, Yoshitada Sakai","doi":"10.20965/jaciii.2023.p0848","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0848","url":null,"abstract":"Articulation disorder is a condition in which the mouth, tongue, vocal cords, and other parts of the body that play an important role in producing voice are damaged, resulting in the inability to produce sound. To diagnose articulation disorders, the movement and shape of each organ concerned with pronunciation are examined. If necessary, the underlying disease or disorder should be managed properly. In it, a speech therapist tests your pronunciation. The observation of conversation and the examination of the pronunciation of each syllable are used to distinguish between mistakes and the degree of articulation disorder. However, these processes are time consuming and labor intensive and are subjective judgments by experts. Therefore, it is important to investigate the characteristics of vocal signals by acoustic analysis of speech objectively. In this study, we focused on fluctuations in the period and amplitude of speech signals and predicted a model for detecting abnormal articulations using fluctuation measurement of the voice data in six healthy subjects and nine patients with an articulation disorder. We used inverse probability of treatment weighting to match the variability for the two groups using the inverse of propensity scores. As the results, the classification performance area under the curve was 0.781 (sensitivity: 0.781, specificity: 0.680) for healthy subjects and patients. We conclude that acoustic analyzing techniques are useful for diagnosing and treating articulation disorders.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264985","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0948
Xiaodong Feng, Lingzhi Yi, Ning Liu, Xieyi Gao, Weiwei Liu, Bin Wang
Edge computing is a new computing method, and task scheduling is challenging work. Using edge computing in intelligent buildings for managing smart home devices has gained popularity because it can reduce the delay and network congestion brought by cloud computing. Edge computing has the advantage of fast response speeds, but its computing capacity is limited. To solve this practical problem, a system framework of collaborative cloud and edge computing is constructed for intelligent buildings. First, the communication time, task completion time, and CPU energy consumption are considered comprehensively, and a mathematical model of the system is developed. Considering the compute-intensity task, the splitting ratio is determined for tasks to achieve the collaboration of cloud computing and edge computing. Then, the search mechanism of a single gene mutation in the genetic algorithm (GA) is introduced to compensate for the defects of the salp swarm algorithm (SSA), while focusing on the search ability and optimization efficiency. Finally, the proposed strategy is theoretically analyzed and experimentally evaluated. The simulation results show that the hybrid algorithm of SSA-GA has better performance than other algorithms, and the proposed collaborative cloud and edge computing task scheduling strategy demonstrated a lower delay and makespan.
{"title":"An Efficient Scheduling Strategy for Collaborative Cloud and Edge Computing in System of Intelligent Buildings","authors":"Xiaodong Feng, Lingzhi Yi, Ning Liu, Xieyi Gao, Weiwei Liu, Bin Wang","doi":"10.20965/jaciii.2023.p0948","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0948","url":null,"abstract":"Edge computing is a new computing method, and task scheduling is challenging work. Using edge computing in intelligent buildings for managing smart home devices has gained popularity because it can reduce the delay and network congestion brought by cloud computing. Edge computing has the advantage of fast response speeds, but its computing capacity is limited. To solve this practical problem, a system framework of collaborative cloud and edge computing is constructed for intelligent buildings. First, the communication time, task completion time, and CPU energy consumption are considered comprehensively, and a mathematical model of the system is developed. Considering the compute-intensity task, the splitting ratio is determined for tasks to achieve the collaboration of cloud computing and edge computing. Then, the search mechanism of a single gene mutation in the genetic algorithm (GA) is introduced to compensate for the defects of the salp swarm algorithm (SSA), while focusing on the search ability and optimization efficiency. Finally, the proposed strategy is theoretically analyzed and experimentally evaluated. The simulation results show that the hybrid algorithm of SSA-GA has better performance than other algorithms, and the proposed collaborative cloud and edge computing task scheduling strategy demonstrated a lower delay and makespan.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264991","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0896
Jiayi Liao, Jing Zheng
With the increasing size of financial assets and the complexity of monetary patterns, countries around the world are gradually becoming more transparent in their monetary policies, using central bank monetary policy communication as a new type of monetary policy instrument. To measure central bank communication more accurately, this paper proposes a dynamic topic model, LDA-BP, based on branching processes, to construct the central bank communication index. At the same time, this paper does four things: it uses the constructed communication index as a proxy variable for the new monetary policy instrument; it builds a TVP-FAVAR model that can extract potential macroeconomic information from many variables, and its time-varying nature can better reflect the dynamic regulatory effect of monetary policy; it constructs a three-dimensional impulse response diagram; and it conducts a systematic analysis of macroeconomic impact. The experimental results of demonstrate its effectiveness on central bank monetary policy communication, as it captures timely information about conventional monetary policy instruments and immediately responds to changes in interest rates and money supply. All three monetary policy instruments are effective in smoothing output volatility, with monetary policy communication having a longer-term impact on the macroeconomy.
{"title":"Research on the Measurement of Central Bank Communication Index and its Impact on the Macroeconomy","authors":"Jiayi Liao, Jing Zheng","doi":"10.20965/jaciii.2023.p0896","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0896","url":null,"abstract":"With the increasing size of financial assets and the complexity of monetary patterns, countries around the world are gradually becoming more transparent in their monetary policies, using central bank monetary policy communication as a new type of monetary policy instrument. To measure central bank communication more accurately, this paper proposes a dynamic topic model, LDA-BP, based on branching processes, to construct the central bank communication index. At the same time, this paper does four things: it uses the constructed communication index as a proxy variable for the new monetary policy instrument; it builds a TVP-FAVAR model that can extract potential macroeconomic information from many variables, and its time-varying nature can better reflect the dynamic regulatory effect of monetary policy; it constructs a three-dimensional impulse response diagram; and it conducts a systematic analysis of macroeconomic impact. The experimental results of demonstrate its effectiveness on central bank monetary policy communication, as it captures timely information about conventional monetary policy instruments and immediately responds to changes in interest rates and money supply. All three monetary policy instruments are effective in smoothing output volatility, with monetary policy communication having a longer-term impact on the macroeconomy.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136264997","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0876
Amina Batool, Yaping Dai, Hongbin Ma, Sijie Yin
In the digital world, automatic component classification is becoming increasingly essential for industrial and logistics applications. The ability to automatically classify various machine parts, such as bolts, nuts, locating pins, bearings, plugs, springs, and washers; using computer vision is challenging for image-based object recognition and classification. Despite varying shapes and classes, components are difficult to distinguish when they appear identical in several ways–particularly in images. This paper proposes identifying machine parts by a deep feature fusion classification model (DFFCM)-variance based designed through the convolutional neural network (CNN), by extracting features and forwarding them to an AdaBoost classifier. DFFCM-v extracts multilayered features from input images, including precise information from image edges, and processes them based on variance. The resulting deep vectors with higher variance are fused using weighted feature fusion to differentiate similar images and used as input to the ensemble AdaBoost classifier for classification. The proposed DFFCM-variance approach achieves the highest accuracy of 99.52% with 341,799 trainable parameters compared with the existing CNN and one-shot learning models, demonstrating its effectiveness in distinguishing similar images of machine components and accurately classifying them.
{"title":"Deep Feature Fusion Classification Model for Identifying Machine Parts","authors":"Amina Batool, Yaping Dai, Hongbin Ma, Sijie Yin","doi":"10.20965/jaciii.2023.p0876","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0876","url":null,"abstract":"In the digital world, automatic component classification is becoming increasingly essential for industrial and logistics applications. The ability to automatically classify various machine parts, such as bolts, nuts, locating pins, bearings, plugs, springs, and washers; using computer vision is challenging for image-based object recognition and classification. Despite varying shapes and classes, components are difficult to distinguish when they appear identical in several ways–particularly in images. This paper proposes identifying machine parts by a deep feature fusion classification model (DFFCM)-variance based designed through the convolutional neural network (CNN), by extracting features and forwarding them to an AdaBoost classifier. DFFCM-v extracts multilayered features from input images, including precise information from image edges, and processes them based on variance. The resulting deep vectors with higher variance are fused using weighted feature fusion to differentiate similar images and used as input to the ensemble AdaBoost classifier for classification. The proposed DFFCM-variance approach achieves the highest accuracy of 99.52% with 341,799 trainable parameters compared with the existing CNN and one-shot learning models, demonstrating its effectiveness in distinguishing similar images of machine components and accurately classifying them.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136265003","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 : 2023-09-20DOI: 10.20965/jaciii.2023.p0771
Gengwang Liu, Yue Yang, Wanting Mo, Wentao Gu, Rihan Wang
The private placement of A-shares gained momentum with the release of the Administrative Measures for Securities Issuance of Listed Companies in 2006. This led to enhanced research on the impact of private placement on stock prices. In 2012, the Chinese government relaxed the requirements for directed issuance of listed companies, resulting in a surge of directed issuance since then. This study uses a sample of listed companies that conducted private placements in the A-share market between 2013 and 2021, to analyze the impact of investor sentiment on stock price differences after private placements from the perspective of short and long-term excess returns. This study constructs the non-main investor sentiment of individual stocks using high-frequency tick data of individual stocks and explores the relationship between this stock price anomaly and investor sentiment using multiple regression analysis. The results show a positive short-term announcement effect of A-share private placements, with the excess return rate occurring mainly before the plan announcement date. The stock price difference from the plan announcement date to ten trading days thereafter has a significantly negative relationship with the excess return rate. Furthermore, investor sentiment in private placements may negatively affect long-term stock performance. This study suggests that this phenomenon is caused by higher investor sentiment pushing stock prices upward in the short term, causing them to deviate from fundamentals, creating mispricing, and then driving them back to fundamentals, with information disclosure. After controlling for the severity of information disclosure, the effect of investor sentiment on long-term stock price performance becomes insignificant.
{"title":"Private Placement, Investor Sentiment, and Stock Price Anomaly","authors":"Gengwang Liu, Yue Yang, Wanting Mo, Wentao Gu, Rihan Wang","doi":"10.20965/jaciii.2023.p0771","DOIUrl":"https://doi.org/10.20965/jaciii.2023.p0771","url":null,"abstract":"The private placement of A-shares gained momentum with the release of the Administrative Measures for Securities Issuance of Listed Companies in 2006. This led to enhanced research on the impact of private placement on stock prices. In 2012, the Chinese government relaxed the requirements for directed issuance of listed companies, resulting in a surge of directed issuance since then. This study uses a sample of listed companies that conducted private placements in the A-share market between 2013 and 2021, to analyze the impact of investor sentiment on stock price differences after private placements from the perspective of short and long-term excess returns. This study constructs the non-main investor sentiment of individual stocks using high-frequency tick data of individual stocks and explores the relationship between this stock price anomaly and investor sentiment using multiple regression analysis. The results show a positive short-term announcement effect of A-share private placements, with the excess return rate occurring mainly before the plan announcement date. The stock price difference from the plan announcement date to ten trading days thereafter has a significantly negative relationship with the excess return rate. Furthermore, investor sentiment in private placements may negatively affect long-term stock performance. This study suggests that this phenomenon is caused by higher investor sentiment pushing stock prices upward in the short term, causing them to deviate from fundamentals, creating mispricing, and then driving them back to fundamentals, with information disclosure. After controlling for the severity of information disclosure, the effect of investor sentiment on long-term stock price performance becomes insignificant.","PeriodicalId":45921,"journal":{"name":"Journal of Advanced Computational Intelligence and Intelligent Informatics","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136263556","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：