Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669031
Akm Ashiquzzaman, L. Ma, Sangwoo Kim, Dongsu Lee, Tai-Won Um, Jinsul Kim
Deeplearning based image classifier is getting improved day by day. The network architecture is also increasing with the accuracy. But the bigger size and resource intensive training makes this model impractical to deploy in IoT based computational units. IoT has limited resources and reckoning power. So smaller network with same accuracy is highly priced for IoT based application deployment. In this study, convolutional deeplearning neural network and how pruning filters without compromising accuracy was studied. Efficient result was achieved from the pruned deeplearning neural network. the model was configured in the experiments by pruning the filter based on absolution position of zeros value based filter ranking. SCADA applications with intelligent component to detect data abnormality and remote sensing also required neural network applications. Using compact memory efficient module in such machines will also give proper validation in such applications in real time. In the end, proposed method for the pruned network delivered same accuracy with reduced size and thus archiving memory and computation for small sized application.
{"title":"Compacting Deep Neural Networks for Light Weight IoT & SCADA Based Applications with Node Pruning","authors":"Akm Ashiquzzaman, L. Ma, Sangwoo Kim, Dongsu Lee, Tai-Won Um, Jinsul Kim","doi":"10.1109/ICAIIC.2019.8669031","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669031","url":null,"abstract":"Deeplearning based image classifier is getting improved day by day. The network architecture is also increasing with the accuracy. But the bigger size and resource intensive training makes this model impractical to deploy in IoT based computational units. IoT has limited resources and reckoning power. So smaller network with same accuracy is highly priced for IoT based application deployment. In this study, convolutional deeplearning neural network and how pruning filters without compromising accuracy was studied. Efficient result was achieved from the pruned deeplearning neural network. the model was configured in the experiments by pruning the filter based on absolution position of zeros value based filter ranking. SCADA applications with intelligent component to detect data abnormality and remote sensing also required neural network applications. Using compact memory efficient module in such machines will also give proper validation in such applications in real time. In the end, proposed method for the pruned network delivered same accuracy with reduced size and thus archiving memory and computation for small sized application.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132085941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669045
Hyun-kyo Lim, Ju-Bong Kim, Joo-Seong Heo, Kwihoon Kim, Yong-Geun Hong, Youn-Hee Han
Recently, the advent of many network applications has led to a tremendous amount of network traffic. A network operator must provide quality of service for each application on the network. To accomplish this goal, various studies have focused on accurately classifying application network traffic. Network management requires technology to classify network traffic without the intervention of the network operator. In this study, we generate packet-based datasets through our own network traffic pre-processing. We train five deep learning models using the convolutional neural network (CNN) and residual network (ResNet) to perform network traffic classification. Finally, we analyze the network traffic classification performance of packet-based datasets using the f1 score of the CNN and ResNet deep learning models, and demonstrate their effectiveness.
{"title":"Packet-based Network Traffic Classification Using Deep Learning","authors":"Hyun-kyo Lim, Ju-Bong Kim, Joo-Seong Heo, Kwihoon Kim, Yong-Geun Hong, Youn-Hee Han","doi":"10.1109/ICAIIC.2019.8669045","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669045","url":null,"abstract":"Recently, the advent of many network applications has led to a tremendous amount of network traffic. A network operator must provide quality of service for each application on the network. To accomplish this goal, various studies have focused on accurately classifying application network traffic. Network management requires technology to classify network traffic without the intervention of the network operator. In this study, we generate packet-based datasets through our own network traffic pre-processing. We train five deep learning models using the convolutional neural network (CNN) and residual network (ResNet) to perform network traffic classification. Finally, we analyze the network traffic classification performance of packet-based datasets using the f1 score of the CNN and ResNet deep learning models, and demonstrate their effectiveness.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133002102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669078
Syed Maaz Shahid, BaekDu Jo, Sunghoon Ko, Sungoh Kwon
In this paper, we propose an engine load classification algorithm using torque data in the crank-angle domain. Engine cylinder operation is different at different engine loads. Engine load information helps to predict the chances or understanding the behavior of a malfunction in engine operation. Hence, we developed an engine load classifier based on signal processing and using an artificial neural network. To that end, we use a magnetic pickup sensor to extract a four-stroke V-type diesel engine's operational information. The pickup sensor's signals are converted to the crank-angle domain (CAD) signal and CAD signals are used in conjunction with the proposed classifier to classify the engine load. For verification, we considered two engine loads (100% and 75%) for a V-type 12-cylinder diesel engine. The proposed algorithm classifies these engine loads with 100% efficiency.
{"title":"Neural Network-based Classification for Engine Load","authors":"Syed Maaz Shahid, BaekDu Jo, Sunghoon Ko, Sungoh Kwon","doi":"10.1109/ICAIIC.2019.8669078","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669078","url":null,"abstract":"In this paper, we propose an engine load classification algorithm using torque data in the crank-angle domain. Engine cylinder operation is different at different engine loads. Engine load information helps to predict the chances or understanding the behavior of a malfunction in engine operation. Hence, we developed an engine load classifier based on signal processing and using an artificial neural network. To that end, we use a magnetic pickup sensor to extract a four-stroke V-type diesel engine's operational information. The pickup sensor's signals are converted to the crank-angle domain (CAD) signal and CAD signals are used in conjunction with the proposed classifier to classify the engine load. For verification, we considered two engine loads (100% and 75%) for a V-type 12-cylinder diesel engine. The proposed algorithm classifies these engine loads with 100% efficiency.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133310064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8668968
S. Narieda, Daiki Cho, K. Umebayashi, H. Naruse
This paper presents weight computation techniques for spectrum sensing based on a cyclic autocorrelation function (CAF) shared diversity combining. We had reported that the performance of signal detection can be improved by the weight factor obtained from time-averaged of the CAF values, and the technique is based on cyclostationary detection based spectrum sensing. In the technique, time-averaged CAFs are used to extract a channel state information and compute a weight factor for the spectrum sensing based on the CAFs. However, the weight factor also includes the CAFs computed by purely additive white Gaussian noise, and the performance of signal detection degrades. In this paper, only the CAFs when it is judged that a primary user is presence are employed to obtain the time-averaged CAF. The presented results show that the performance of signal detection can be improved as compared with the conventional weight computation technique.
{"title":"Statistics Shared CAF Diversity Combining Based Sensing Using Weight Computation Technique","authors":"S. Narieda, Daiki Cho, K. Umebayashi, H. Naruse","doi":"10.1109/ICAIIC.2019.8668968","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8668968","url":null,"abstract":"This paper presents weight computation techniques for spectrum sensing based on a cyclic autocorrelation function (CAF) shared diversity combining. We had reported that the performance of signal detection can be improved by the weight factor obtained from time-averaged of the CAF values, and the technique is based on cyclostationary detection based spectrum sensing. In the technique, time-averaged CAFs are used to extract a channel state information and compute a weight factor for the spectrum sensing based on the CAFs. However, the weight factor also includes the CAFs computed by purely additive white Gaussian noise, and the performance of signal detection degrades. In this paper, only the CAFs when it is judged that a primary user is presence are employed to obtain the time-averaged CAF. The presented results show that the performance of signal detection can be improved as compared with the conventional weight computation technique.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114081047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8668997
Kiho Choi, Hyeongrae Kim, Daejin Park, Jeonghun Cho
Signature-based control flow monitoring is a representative technique for detecting control flow errors in run time. However, it is very inefficient and time consuming to manually insert the monitoring code into a monitor-target application. In particular, for performance improvements of control-flow monitoring, implementing a monitoring code that operates in multi-thread makes things more complicated. In this paper, we propose an automatic code-generation framework that automatically translate an application into the control-flow monitorable application. In the proposed framework, the applied technique for control-flow monitoring is based on separate signature-based control-flow monitoring (SSCFM) technique that is able to expect performance improvements in multi-threaded or multi-core environments by separating the signature update and the signature verification on the thread level. The proposed framework automatically analyzes a monitor-target application and generates a SSCFM-applied application based on the analysis results. We anticipate that our automatic multi-thread code generation framework for control flow monitoring lessens the burden in runtime control-flow monitoring field.
{"title":"Automatic Multi-Thread Code Generation for Monitoring Signature-based Control Flow","authors":"Kiho Choi, Hyeongrae Kim, Daejin Park, Jeonghun Cho","doi":"10.1109/ICAIIC.2019.8668997","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8668997","url":null,"abstract":"Signature-based control flow monitoring is a representative technique for detecting control flow errors in run time. However, it is very inefficient and time consuming to manually insert the monitoring code into a monitor-target application. In particular, for performance improvements of control-flow monitoring, implementing a monitoring code that operates in multi-thread makes things more complicated. In this paper, we propose an automatic code-generation framework that automatically translate an application into the control-flow monitorable application. In the proposed framework, the applied technique for control-flow monitoring is based on separate signature-based control-flow monitoring (SSCFM) technique that is able to expect performance improvements in multi-threaded or multi-core environments by separating the signature update and the signature verification on the thread level. The proposed framework automatically analyzes a monitor-target application and generates a SSCFM-applied application based on the analysis results. We anticipate that our automatic multi-thread code generation framework for control flow monitoring lessens the burden in runtime control-flow monitoring field.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121056333","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669085
M. Mohammadi, Mursal Dawodi, Tomohisa Wada, Nadira Ahmadi
With huge amount of data in diverse technological areas, and generating such kinds of data rapidly, it needs for proper usage; therefore, Data Mining has emerged. Data Mining can extract prominent knowledge from customary data that can attract attention of people to it which is meaningful information. Regarding this concept that data can be generated rapidly every day or even every moment, data need to take under process for offering better valuable information. Data of educational areas is more that belongs to students, and it's all right a good basis for commence of applying Data Mining. In this paper the focus is on how to use Data Mining techniques to discover information in student`s raw data and different algorithms such as KNN, Naïve Bayes, and Decision Tree are implemented.
{"title":"Comparative study of supervised learning algorithms for student performance prediction","authors":"M. Mohammadi, Mursal Dawodi, Tomohisa Wada, Nadira Ahmadi","doi":"10.1109/ICAIIC.2019.8669085","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669085","url":null,"abstract":"With huge amount of data in diverse technological areas, and generating such kinds of data rapidly, it needs for proper usage; therefore, Data Mining has emerged. Data Mining can extract prominent knowledge from customary data that can attract attention of people to it which is meaningful information. Regarding this concept that data can be generated rapidly every day or even every moment, data need to take under process for offering better valuable information. Data of educational areas is more that belongs to students, and it's all right a good basis for commence of applying Data Mining. In this paper the focus is on how to use Data Mining techniques to discover information in student`s raw data and different algorithms such as KNN, Naïve Bayes, and Decision Tree are implemented.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"87 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122951677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8668972
Abduladhim Ashtaiwi
Human deaths, injuries caused by road crashes have tremendous impacts on individuals, families, and societies. Economically, it causes financial burden on countries as, on average, they loss of 3% of their Gross Domestic Product (GDP). Many driving assistant techniques, embedded in several vehicles, are helping drivers to avoid car crashes by giving them early warning message. In this work, An Intelligent Road Crashes Avoidance (IRCA) system which adopts the Artificial Neural Network (ANN) and Decision Tree (DT) algorithms is proposed. The prediction model of IRCA is trained using big dataset composed of 1.6 million rows (car accidents) and 23 features (information) spanning over 14 years of data collection by United Kingdom (UK). With prediction accuracy of 72% for ANN and 74% for TD algorithms, IRCA system can predict car crash risk levels for 941 districts of UK. IRCA system can be exploited either in human-driven or in self-driving cars. The prediction accuracy can further be improved by training on new collected dataset with less missing data and outliers.
{"title":"Intelligent Road Crashes Avoidance System","authors":"Abduladhim Ashtaiwi","doi":"10.1109/ICAIIC.2019.8668972","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8668972","url":null,"abstract":"Human deaths, injuries caused by road crashes have tremendous impacts on individuals, families, and societies. Economically, it causes financial burden on countries as, on average, they loss of 3% of their Gross Domestic Product (GDP). Many driving assistant techniques, embedded in several vehicles, are helping drivers to avoid car crashes by giving them early warning message. In this work, An Intelligent Road Crashes Avoidance (IRCA) system which adopts the Artificial Neural Network (ANN) and Decision Tree (DT) algorithms is proposed. The prediction model of IRCA is trained using big dataset composed of 1.6 million rows (car accidents) and 23 features (information) spanning over 14 years of data collection by United Kingdom (UK). With prediction accuracy of 72% for ANN and 74% for TD algorithms, IRCA system can predict car crash risk levels for 941 districts of UK. IRCA system can be exploited either in human-driven or in self-driving cars. The prediction accuracy can further be improved by training on new collected dataset with less missing data and outliers.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121845756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669064
Huy Nguyen, Minh Duc Thieu, Tung Lam Pham, Hoan Nguyen, Y. Jang
Unlike visible light communication (VLC) using photo detector, optical camera communication (OCC) employs an image sensor as the receiver. In this paper, we discuss the parameter of rolling shutter camera (focal length, FOV, rolling rate, frame rate) which impact to the data rate of optical camera communication system. Besides that, we proposed the suitable parameter of cameras for optical camera communication system to enhance the higher performance of data transmission.
{"title":"The Impact of Camera Parameters on Optical Camera Communication","authors":"Huy Nguyen, Minh Duc Thieu, Tung Lam Pham, Hoan Nguyen, Y. Jang","doi":"10.1109/ICAIIC.2019.8669064","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669064","url":null,"abstract":"Unlike visible light communication (VLC) using photo detector, optical camera communication (OCC) employs an image sensor as the receiver. In this paper, we discuss the parameter of rolling shutter camera (focal length, FOV, rolling rate, frame rate) which impact to the data rate of optical camera communication system. Besides that, we proposed the suitable parameter of cameras for optical camera communication system to enhance the higher performance of data transmission.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116867733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669087
Jhe-Wei Lin, Yunwen Gao, Rong-Guey Chang
The sequence transformer models are based on complex recurrent neural network or convolutional networks that include an encoder and a decoder. High-accuracy models are usually represented by used connect the encoder and decoder through an attention mechanism. Story generation is an important thing. If we can let computers learn the ability of story-telling, computers can help people do more things. Actually, the squence2squence model combine attention mechanism is being used to Chinese poetry generation. However, it difficult to apply in Chinese story generation, because there are some rules in Chinese poetry generation. Therefore, we trying to use 1372 human-labeled summarization of paragraphs from a classic novel named “Demi-Gods and Semi-Devils” (天龍八部) to train the transformer network. In our experiment, we use FastText to combine Demi-Gods and Semi-Devils Dataset and A Large Scale Chinese Short Text Summarization Dataset to be input data. In addition, we got a lower loss rate by using two layer of self-attention mechanism.
{"title":"Chinese Story Generation with FastText Transformer Network","authors":"Jhe-Wei Lin, Yunwen Gao, Rong-Guey Chang","doi":"10.1109/ICAIIC.2019.8669087","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669087","url":null,"abstract":"The sequence transformer models are based on complex recurrent neural network or convolutional networks that include an encoder and a decoder. High-accuracy models are usually represented by used connect the encoder and decoder through an attention mechanism. Story generation is an important thing. If we can let computers learn the ability of story-telling, computers can help people do more things. Actually, the squence2squence model combine attention mechanism is being used to Chinese poetry generation. However, it difficult to apply in Chinese story generation, because there are some rules in Chinese poetry generation. Therefore, we trying to use 1372 human-labeled summarization of paragraphs from a classic novel named “Demi-Gods and Semi-Devils” (天龍八部) to train the transformer network. In our experiment, we use FastText to combine Demi-Gods and Semi-Devils Dataset and A Large Scale Chinese Short Text Summarization Dataset to be input data. In addition, we got a lower loss rate by using two layer of self-attention mechanism.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122146682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-02-01DOI: 10.1109/ICAIIC.2019.8669035
Yuki Watanabe, Reiji Suzumura, Shogo Matsuno, M. Ohyama
The physical activity is important context information to define and understand the user’s situation in real time and in detail. Therefore, we developed a context-aware function using the activity recognition and showed that it is possible to provide more appropriate support according to the user’s situation. In this study, we first constructed a model by applying machine learning to data sensed by a smartphone in order to predict the physical activity of the user. In the experiment, high accuracy of 97.6% was obtained by using the model. Next, we developed three functions using the activity recognition. These functions predict the physical activity of user in real time. In addition, user support is performed according to the predicted physical activity. In the experiment using developed functions, it is confirmed that these functions worked correctly in real-world conditions.
{"title":"Investigation of Context-aware System Using Activity Recognition","authors":"Yuki Watanabe, Reiji Suzumura, Shogo Matsuno, M. Ohyama","doi":"10.1109/ICAIIC.2019.8669035","DOIUrl":"https://doi.org/10.1109/ICAIIC.2019.8669035","url":null,"abstract":"The physical activity is important context information to define and understand the user’s situation in real time and in detail. Therefore, we developed a context-aware function using the activity recognition and showed that it is possible to provide more appropriate support according to the user’s situation. In this study, we first constructed a model by applying machine learning to data sensed by a smartphone in order to predict the physical activity of the user. In the experiment, high accuracy of 97.6% was obtained by using the model. Next, we developed three functions using the activity recognition. These functions predict the physical activity of user in real time. In addition, user support is performed according to the predicted physical activity. In the experiment using developed functions, it is confirmed that these functions worked correctly in real-world conditions.","PeriodicalId":273383,"journal":{"name":"2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122307429","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：