Pub Date : 2019-07-10DOI: 10.1109/JCSSE.2019.8864178
Khushbu Gupta, Ratchainant Thammasudjarit, A. Thakkinstian
Clinical researches and practitioners require data extracted from CT scan reports but most of them are in unstructured data format, which are not ready to analysis. Furthermore, a lag of annotated data makes data extraction more difficult to apply natural language processing techniques to convert unstructured data to be structured data. This study is therefore conducted to apply an automated engine employing topic modeling combined with lexicon and syntactic rule-based approach to extract clinical information from CT scan reports. This prototype shows promising results for constructing clinical datasets for further clinical researches.
{"title":"A Hybrid Engine for Clinical Information Extraction from Radiology Reports","authors":"Khushbu Gupta, Ratchainant Thammasudjarit, A. Thakkinstian","doi":"10.1109/JCSSE.2019.8864178","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864178","url":null,"abstract":"Clinical researches and practitioners require data extracted from CT scan reports but most of them are in unstructured data format, which are not ready to analysis. Furthermore, a lag of annotated data makes data extraction more difficult to apply natural language processing techniques to convert unstructured data to be structured data. This study is therefore conducted to apply an automated engine employing topic modeling combined with lexicon and syntactic rule-based approach to extract clinical information from CT scan reports. This prototype shows promising results for constructing clinical datasets for further clinical researches.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115340412","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 : 2019-07-10DOI: 10.1109/JCSSE.2019.8864156
Rasa Bhattarai, M. Phothisonothai
Uniqueness in the analysis pattern of objects by individual humans has a profound impact on the study of their visual learning and behavior. Eye movement patterns have been effectively emerging as a biometric based key for security systems, product recognition patterns, user identifications, as well as medical research purposes. The modern eye tracking systems are non-invasive and financially affordable. Therefore, in this paper, we proposed eye-tracking based visualizations and metrics analysis for individual eye movement patterns collected during any kinds of activities depending on the scope of the our experimental paradigms. Individuals can be aware of their own performances during certain task and improve upon their weak areas. The objective of the paper is to utilize the important visual metrics obtained from fixation, saccades and face recognition and use them to analyze for individual categorization. The obtained results shown that the specific features and patterns can be extracted the viewing aspect of individual subjects using naive Bayes classifier. We were successfully able to predict the individual eye movements with an accuracy of 90.22%.
{"title":"Eye-Tracking Based Visualizations and Metrics Analysis for Individual Eye Movement Patterns","authors":"Rasa Bhattarai, M. Phothisonothai","doi":"10.1109/JCSSE.2019.8864156","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864156","url":null,"abstract":"Uniqueness in the analysis pattern of objects by individual humans has a profound impact on the study of their visual learning and behavior. Eye movement patterns have been effectively emerging as a biometric based key for security systems, product recognition patterns, user identifications, as well as medical research purposes. The modern eye tracking systems are non-invasive and financially affordable. Therefore, in this paper, we proposed eye-tracking based visualizations and metrics analysis for individual eye movement patterns collected during any kinds of activities depending on the scope of the our experimental paradigms. Individuals can be aware of their own performances during certain task and improve upon their weak areas. The objective of the paper is to utilize the important visual metrics obtained from fixation, saccades and face recognition and use them to analyze for individual categorization. The obtained results shown that the specific features and patterns can be extracted the viewing aspect of individual subjects using naive Bayes classifier. We were successfully able to predict the individual eye movements with an accuracy of 90.22%.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131692392","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 : 2019-07-10DOI: 10.1109/JCSSE.2019.8864167
Duangduen Asavasuthirakul, Sittha Saisawan, A. Harfield, Prasert Wiangsukphaiboon
Geographical positioning is indispensable in fields such as agriculture automation, mapping, and land surveying. Real-time kinematic (RTK) is a technique to enhance the positional accuracy of global navigation satellite systems (GNSS) to centimeter level precision. However, professional-grade RTK GNSS devices and their associated commercial tools are expensive and require expertise to operate. Thus, the authors have developed a low-cost RTK GNSS receiver “Pantai” together with its cloud control center, called “RTK Control Center”. Pantai supports satellite signals from multi-constellations and can communicate across existing data networks (3G+/Wi-Fi) to provide real-time positioning information and to receive operation commands from remote users via the control center. In this paper, we describe the architecture for Pantai and RTK Control Center, and we undertake two experiments to verify the accuracy and reliability of the system. The results show that the system can measure the position of a rover with centimeter level accuracy.
{"title":"A Low-Cost RTK GNSS Receiver with Cloud-Based Control Center Application","authors":"Duangduen Asavasuthirakul, Sittha Saisawan, A. Harfield, Prasert Wiangsukphaiboon","doi":"10.1109/JCSSE.2019.8864167","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864167","url":null,"abstract":"Geographical positioning is indispensable in fields such as agriculture automation, mapping, and land surveying. Real-time kinematic (RTK) is a technique to enhance the positional accuracy of global navigation satellite systems (GNSS) to centimeter level precision. However, professional-grade RTK GNSS devices and their associated commercial tools are expensive and require expertise to operate. Thus, the authors have developed a low-cost RTK GNSS receiver “Pantai” together with its cloud control center, called “RTK Control Center”. Pantai supports satellite signals from multi-constellations and can communicate across existing data networks (3G+/Wi-Fi) to provide real-time positioning information and to receive operation commands from remote users via the control center. In this paper, we describe the architecture for Pantai and RTK Control Center, and we undertake two experiments to verify the accuracy and reliability of the system. The results show that the system can measure the position of a rover with centimeter level accuracy.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126656454","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864220
Annupan Rodtook, Khwunta Kirimasthong
A number of popular methods for segmentation are based on a generalized gradient vector flow (GGVF) snakes. One of the most successful recent extensions of this idea is the adaptive diffusion flow (ADF) snake. However, the good choice of the control parameters of the ADF such as the Gaussian smoothing coefficients and the weights associated with harmonic hypersurface functional and the infinity Laplacian are hard to select. In turn, the wrong choice often leads to inappropriate results. In this paper, we propose a new method to select the control parameters of ADF based on the vector field analysis (VFA). The experimental results obtained on a set of 40 US images of breast tumor show that the ADF combined with the VFA outperforms the original ADF.
{"title":"Circular Vector Field Analysis for the Adaptive Diffusion Flow Snakes Applied to Ultrasound Images of Breast Cancer","authors":"Annupan Rodtook, Khwunta Kirimasthong","doi":"10.1109/JCSSE.2019.8864220","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864220","url":null,"abstract":"A number of popular methods for segmentation are based on a generalized gradient vector flow (GGVF) snakes. One of the most successful recent extensions of this idea is the adaptive diffusion flow (ADF) snake. However, the good choice of the control parameters of the ADF such as the Gaussian smoothing coefficients and the weights associated with harmonic hypersurface functional and the infinity Laplacian are hard to select. In turn, the wrong choice often leads to inappropriate results. In this paper, we propose a new method to select the control parameters of ADF based on the vector field analysis (VFA). The experimental results obtained on a set of 40 US images of breast tumor show that the ADF combined with the VFA outperforms the original ADF.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125344681","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864198
Jumpol Yaothanee, K. Chanchio
A cluster of virtual machines can be used to execute parallel applications in Cloud Computing environments. However, the cloud infrastructure may fail at any time for a variety of reasons. Although a coordinated checkpointing capability at the hypervisor level is highly transparent to parallel applications, existing solutions still suffer from excessive checkpoint time and downtime. They also cause significant application execution delays due to packet loss. This paper introduces IMVCCR, a novel in-memory Checkpoint-Restart mechanism for a virtual cluster. IMVCCR consists of a framework that performs coordinated checkpointing for the entire cluster. It reduces checkpoint time and downtime by applying live migration and using main memory as transient checkpoint storage. IMVCCR also uses an efficient synchronization mechanism to reduce packet loss. Preliminary experiments show that IMVCCR generates very low checkpoint times and downtimes. It also incurs low overheads in the total execution time of parallel applications.
{"title":"An In-Memory Checkpoint-Restart Mechanism for a Cluster of Virtual Machines","authors":"Jumpol Yaothanee, K. Chanchio","doi":"10.1109/JCSSE.2019.8864198","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864198","url":null,"abstract":"A cluster of virtual machines can be used to execute parallel applications in Cloud Computing environments. However, the cloud infrastructure may fail at any time for a variety of reasons. Although a coordinated checkpointing capability at the hypervisor level is highly transparent to parallel applications, existing solutions still suffer from excessive checkpoint time and downtime. They also cause significant application execution delays due to packet loss. This paper introduces IMVCCR, a novel in-memory Checkpoint-Restart mechanism for a virtual cluster. IMVCCR consists of a framework that performs coordinated checkpointing for the entire cluster. It reduces checkpoint time and downtime by applying live migration and using main memory as transient checkpoint storage. IMVCCR also uses an efficient synchronization mechanism to reduce packet loss. Preliminary experiments show that IMVCCR generates very low checkpoint times and downtimes. It also incurs low overheads in the total execution time of parallel applications.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130402836","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864181
Suphamongkol Akkaradamrongrat, Pornpimon Kachamas, S. Sinthupinyo
The problem of imbalanced data can be frequently found in the real-world data. It leads to the bias of classification models, that is, the models predict most samples as major classes which are often the negative class. In this research, text generation techniques were used to generate synthetic minority class samples to make the text dataset balanced. Two text generation methods: the text generation using Markov Chains and the text generation using Long Short-term Memory (LSTM) networks were applied and compared in the term of ability to improve the performance of imbalanced text classification. Our experimental study is based on LSTM networks classifier. Traditional over-sampling technique was also used as baseline. The study investigated our Thai-language advertisement text dataset from Facebook. According to the increase of recall value, applying of these techniques showed the improvement of an ability to create model predicting more positive samples, which are minority samples. It can be found that the Markov Chains technique outperformed traditional over-sampling and text generation using LSTM in majority of the models.
{"title":"Text Generation for Imbalanced Text Classification","authors":"Suphamongkol Akkaradamrongrat, Pornpimon Kachamas, S. Sinthupinyo","doi":"10.1109/JCSSE.2019.8864181","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864181","url":null,"abstract":"The problem of imbalanced data can be frequently found in the real-world data. It leads to the bias of classification models, that is, the models predict most samples as major classes which are often the negative class. In this research, text generation techniques were used to generate synthetic minority class samples to make the text dataset balanced. Two text generation methods: the text generation using Markov Chains and the text generation using Long Short-term Memory (LSTM) networks were applied and compared in the term of ability to improve the performance of imbalanced text classification. Our experimental study is based on LSTM networks classifier. Traditional over-sampling technique was also used as baseline. The study investigated our Thai-language advertisement text dataset from Facebook. According to the increase of recall value, applying of these techniques showed the improvement of an ability to create model predicting more positive samples, which are minority samples. It can be found that the Markov Chains technique outperformed traditional over-sampling and text generation using LSTM in majority of the models.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127635932","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864168
Thongtham Chongmesuk, Vishnu Kotrajaras
Existing quest generation systems that use structural rules have an important limitation. A quest generated by such systems are not guaranteed to be flexible such that players can finish the quest in a variety of ways. In this paper, we extend such systems by replacing action-based quests with game state-based quests. A quest generated from our system is filtered for conflicting scenarios and then analyzed using Prolog to guarantee multiple paths of completion. Ensuring that quests always have multiple ways to complete improves quest quality and players' experience.
{"title":"Multi-Paths Generation for Structural Rule Quests","authors":"Thongtham Chongmesuk, Vishnu Kotrajaras","doi":"10.1109/JCSSE.2019.8864168","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864168","url":null,"abstract":"Existing quest generation systems that use structural rules have an important limitation. A quest generated by such systems are not guaranteed to be flexible such that players can finish the quest in a variety of ways. In this paper, we extend such systems by replacing action-based quests with game state-based quests. A quest generated from our system is filtered for conflicting scenarios and then analyzed using Prolog to guarantee multiple paths of completion. Ensuring that quests always have multiple ways to complete improves quest quality and players' experience.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"220 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133999312","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}
In this paper, we propose a new recommendation method, named “Cross-Category Product with Diversity and Novelty” which uses association rule mining and analytic hierarchy process to recommend a personalized rank list to target users. Long-tail products are also inserted into the list for improving diversity and novelty. The experimental results indicate that our method has higher discounted cumulative gain (DCG), diversity, novelty, and coverage than the research that makes cross-category recommendations on single-criterion rating and the research that uses only collaborative filtering on multi-criteria ratings.
{"title":"Cross-Category Product Recommender System based on Multi-Criteria Rating using Diversity and Novelty Evaluation","authors":"Saranya Maneeroj, Pongsakorn Jirachanchaisiri, Chanisara Suksomjit, Apirom Zatloukal","doi":"10.1109/JCSSE.2019.8864193","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864193","url":null,"abstract":"In this paper, we propose a new recommendation method, named “Cross-Category Product with Diversity and Novelty” which uses association rule mining and analytic hierarchy process to recommend a personalized rank list to target users. Long-tail products are also inserted into the list for improving diversity and novelty. The experimental results indicate that our method has higher discounted cumulative gain (DCG), diversity, novelty, and coverage than the research that makes cross-category recommendations on single-criterion rating and the research that uses only collaborative filtering on multi-criteria ratings.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134287616","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864216
Parinya Preamthaisong, Anucha Auyporntrakool, Phet Aimtongkham, Titaya Sriwuttisap, C. So-In
This research has investigated the probable integration of a hybrid classification model into a Distributed Denial of Service (DDoS) detection scheme for Software-Defined Network (SDN). There are four key modules in our framework: 1) Traffic Generator, 2) SDN Controller, 3. Mininet (Openflow enabled switch), and 4) Alert. To enhance the DDoS detection precision, we also propose the use of Genetic Algorithm (GA) with a combination of Decision Tree (DT), called GA-DT. The implementation is based on Mininet as SDN emulator. To confirm our superiority, we practically used the real-trace of the four recent DDoSs, i.e., TCP SYN Flood, UDP Flooding, ICMP Flooding, and TCPKill, captured from Wireshark, with our hybrid classification against the existing ones including DT, Logistic Regression (LR), Neural network (NN), Self-organizing map (SOM), $k$-nearest neighbors (kNN), Support Vector Machine (SVM), and Random forests (RF). The results show that GA-DT outperforms the others in terms of higher accuracy.
本研究探讨了将混合分类模型集成到软件定义网络(SDN)分布式拒绝服务(DDoS)检测方案中的可能性。在我们的框架中有四个关键模块:1)流量生成器，2)SDN控制器，3。Mininet (Openflow使能开关)，4)Alert。为了提高DDoS检测的精度，我们还提出使用遗传算法(GA)和决策树(DT)的组合，称为GA-DT。该实现基于Mininet作为SDN仿真器。为了证实我们的优势，我们实际使用了从Wireshark捕获的最近四次ddos的真实跟踪，即TCP SYN Flood, UDP Flood, ICMP Flood和TCPKill，并对现有的分类进行了混合分类，包括DT，逻辑回归(LR)，神经网络(NN)，自组织映射(SOM)， k近邻(kNN)，支持向量机(SVM)和随机森林(RF)。结果表明，GA-DT在更高的精度方面优于其他方法。
{"title":"Enhanced DDoS Detection using Hybrid Genetic Algorithm and Decision Tree for SDN","authors":"Parinya Preamthaisong, Anucha Auyporntrakool, Phet Aimtongkham, Titaya Sriwuttisap, C. So-In","doi":"10.1109/JCSSE.2019.8864216","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864216","url":null,"abstract":"This research has investigated the probable integration of a hybrid classification model into a Distributed Denial of Service (DDoS) detection scheme for Software-Defined Network (SDN). There are four key modules in our framework: 1) Traffic Generator, 2) SDN Controller, 3. Mininet (Openflow enabled switch), and 4) Alert. To enhance the DDoS detection precision, we also propose the use of Genetic Algorithm (GA) with a combination of Decision Tree (DT), called GA-DT. The implementation is based on Mininet as SDN emulator. To confirm our superiority, we practically used the real-trace of the four recent DDoSs, i.e., TCP SYN Flood, UDP Flooding, ICMP Flooding, and TCPKill, captured from Wireshark, with our hybrid classification against the existing ones including DT, Logistic Regression (LR), Neural network (NN), Self-organizing map (SOM), $k$-nearest neighbors (kNN), Support Vector Machine (SVM), and Random forests (RF). The results show that GA-DT outperforms the others in terms of higher accuracy.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128937177","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 : 2019-07-01DOI: 10.1109/JCSSE.2019.8864184
A. Khine, W. Wettayaprasit, Jarunee Duangsuwan
Nurse notes often contain subjective information of a patient's health status. However, these have not been widely used to predict clinical outcomes. Advances in natural language processing enable to extract information from unstructured clinical documents such as nurse notes. In this paper, we use TF-IDF representations of nurse notes to predict Intensive Care Unit (ICU) mortality while controlling for other candidate features such as gender, ICU type, age of patient at first admission, SAPS (Simplified Acute Physiology Score) II score, SAPS II probability, polarity and subjectivity scores of each nurse note. We introduce an ensemble of CNN and MLP model to predict 30-day ICU mortality. We apply our model to MIMIC III which is a medical benchmark dataset. We use TF-IDF representation of nurse notes as input to CNN and other ICU features to MLP network. Experimental results demonstrate that proposed ensembled model with nurse notes have higher performance than standalone models.
{"title":"Ensemble CNN and MLP with Nurse Notes for Intensive Care Unit Mortality","authors":"A. Khine, W. Wettayaprasit, Jarunee Duangsuwan","doi":"10.1109/JCSSE.2019.8864184","DOIUrl":"https://doi.org/10.1109/JCSSE.2019.8864184","url":null,"abstract":"Nurse notes often contain subjective information of a patient's health status. However, these have not been widely used to predict clinical outcomes. Advances in natural language processing enable to extract information from unstructured clinical documents such as nurse notes. In this paper, we use TF-IDF representations of nurse notes to predict Intensive Care Unit (ICU) mortality while controlling for other candidate features such as gender, ICU type, age of patient at first admission, SAPS (Simplified Acute Physiology Score) II score, SAPS II probability, polarity and subjectivity scores of each nurse note. We introduce an ensemble of CNN and MLP model to predict 30-day ICU mortality. We apply our model to MIMIC III which is a medical benchmark dataset. We use TF-IDF representation of nurse notes as input to CNN and other ICU features to MLP network. Experimental results demonstrate that proposed ensembled model with nurse notes have higher performance than standalone models.","PeriodicalId":313194,"journal":{"name":"2019 16th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132375782","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：