Pub Date : 2019-10-01DOI: 10.1109/ICITEED.2019.8929975
Fuenglada Manokij, Kanoksri Sarinnapakorn, P. Vateekul
Precipitation prediction is necessary to use in water management, especially in Thailand, it can be applied for various ways, such as flood warning, agriculture planning, etc. There are many prior attempts to forecast rainfall from the rain-gauge station. Some deployed traditional machine learning approaches: ARIMA, k-NN, etc. Recently, deep learning approach has shown promising results in this task. However, the accuracy is still limited since the raining period throughout the year in Thailand is very scarce, so most rainfall amount is zero. In this paper, we propose to cascade two deep learning networks to tackle this problem: one is a classification model to classify whether it rains or not, and the other is a regression model to predict rainfall amount. Our network is a combination of CNN and GRU, where CNN aims to capture relationship between various sensors and GRU aims to capture time-series information. Furthermore, we also perform multi-step forecasting by applying a rolling mechanism that uses the predicted rainfall and involved features to predict the next 6 steps. The experiment was conducted on hourly rainfall dataset for 6 years (2013-2018) provided from the public government sector in Thailand. We use RMSE as performance metric to evaluate three periods of rainfall: Overall, Rain, and non-rain periods and the results show that our cascading model is the winner with only 4.53% in term of RMSE which is the average percentage of difference from over all regions.
{"title":"Forecasting Thailand’s Precipitation with Cascading Model of CNN and GRU","authors":"Fuenglada Manokij, Kanoksri Sarinnapakorn, P. Vateekul","doi":"10.1109/ICITEED.2019.8929975","DOIUrl":"https://doi.org/10.1109/ICITEED.2019.8929975","url":null,"abstract":"Precipitation prediction is necessary to use in water management, especially in Thailand, it can be applied for various ways, such as flood warning, agriculture planning, etc. There are many prior attempts to forecast rainfall from the rain-gauge station. Some deployed traditional machine learning approaches: ARIMA, k-NN, etc. Recently, deep learning approach has shown promising results in this task. However, the accuracy is still limited since the raining period throughout the year in Thailand is very scarce, so most rainfall amount is zero. In this paper, we propose to cascade two deep learning networks to tackle this problem: one is a classification model to classify whether it rains or not, and the other is a regression model to predict rainfall amount. Our network is a combination of CNN and GRU, where CNN aims to capture relationship between various sensors and GRU aims to capture time-series information. Furthermore, we also perform multi-step forecasting by applying a rolling mechanism that uses the predicted rainfall and involved features to predict the next 6 steps. The experiment was conducted on hourly rainfall dataset for 6 years (2013-2018) provided from the public government sector in Thailand. We use RMSE as performance metric to evaluate three periods of rainfall: Overall, Rain, and non-rain periods and the results show that our cascading model is the winner with only 4.53% in term of RMSE which is the average percentage of difference from over all regions.","PeriodicalId":6598,"journal":{"name":"2019 11th International Conference on Information Technology and Electrical Engineering (ICITEE)","volume":"163 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78557633","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 : 2019-10-01DOI: 10.1109/ICITEED.2019.8930003
Sumek Wisayataksin, Panupong Angkasuwan, Y. Ariyakul
Odor-shift keying, a data modulation technique that encodes digital data by varying odor presentation, was proposed in this paper. A multi-channel olfactory display was used as modulator to release multiple odors whose blending ratio representing different digital data. On the demodulator side, an odor sensing system is used to measure the released smells, revert them into electrical form which is then decoded into the original data by using digital signal processing. The proposed technique can enhance data transfer rate of the communication through odor as carrier. A preliminary experiment was conducted to validate the possibility of the concept of varying the presented odors to represent different binary data. Finally, a string was practically modulated and transmitted by using the proposed technique. The demodulation process can be performed successfully and the data transfer rate was doubled from the previous work.
{"title":"4-ary Odor-Shift Keying Using Multi-channel Olfactory Display","authors":"Sumek Wisayataksin, Panupong Angkasuwan, Y. Ariyakul","doi":"10.1109/ICITEED.2019.8930003","DOIUrl":"https://doi.org/10.1109/ICITEED.2019.8930003","url":null,"abstract":"Odor-shift keying, a data modulation technique that encodes digital data by varying odor presentation, was proposed in this paper. A multi-channel olfactory display was used as modulator to release multiple odors whose blending ratio representing different digital data. On the demodulator side, an odor sensing system is used to measure the released smells, revert them into electrical form which is then decoded into the original data by using digital signal processing. The proposed technique can enhance data transfer rate of the communication through odor as carrier. A preliminary experiment was conducted to validate the possibility of the concept of varying the presented odors to represent different binary data. Finally, a string was practically modulated and transmitted by using the proposed technique. The demodulation process can be performed successfully and the data transfer rate was doubled from the previous work.","PeriodicalId":6598,"journal":{"name":"2019 11th International Conference on Information Technology and Electrical Engineering (ICITEE)","volume":"24 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81931933","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 : 2019-10-01DOI: 10.1109/ICITEED.2019.8929995
Hanung Adi Nugroho, A. Fatan D. Marsiano, Khampaserth Xaphakdy, Phounsiri Sihakhom, Eka Legya Frannita, Rizki Nurfauzi, E. Elsa Herdiana Murhandarwati
Malaria has become one of the deadliest diseases in the world. The main method in diagnosing malaria is manually conducted by pathologists using a microscope. This process is time-consuming and prone to error due to human factor. These facts encourage the development of system that consistently yielded more objective results regardless of the condition on the field. In this research work, a novel method to segment infected erythrocytes using threshold and morphological is proposed. The proposed method was tested in a database consisting of 30 images with varying condition. The experimental results showed that the proposed method achieved 96.74 ± 0.7075 %, 76.77 ± 2.1441 %, 99.74 ± 0.1397 %, 97.84 ± 1.2514 % and 96.61 ± 0.8021 % of accuracy, sensitivity, specificity, prediction value positive and prediction value negative, respectively. In conclusion, the proposed method provides a consistent result for segmenting parasite in infected erythrocytes image. This result indicates that this proposed scheme is proper to assist the pathologists in detecting Plasmodium parasites.
{"title":"Multithresholding Approach for Segmenting Plasmodium Parasites","authors":"Hanung Adi Nugroho, A. Fatan D. Marsiano, Khampaserth Xaphakdy, Phounsiri Sihakhom, Eka Legya Frannita, Rizki Nurfauzi, E. Elsa Herdiana Murhandarwati","doi":"10.1109/ICITEED.2019.8929995","DOIUrl":"https://doi.org/10.1109/ICITEED.2019.8929995","url":null,"abstract":"Malaria has become one of the deadliest diseases in the world. The main method in diagnosing malaria is manually conducted by pathologists using a microscope. This process is time-consuming and prone to error due to human factor. These facts encourage the development of system that consistently yielded more objective results regardless of the condition on the field. In this research work, a novel method to segment infected erythrocytes using threshold and morphological is proposed. The proposed method was tested in a database consisting of 30 images with varying condition. The experimental results showed that the proposed method achieved 96.74 ± 0.7075 %, 76.77 ± 2.1441 %, 99.74 ± 0.1397 %, 97.84 ± 1.2514 % and 96.61 ± 0.8021 % of accuracy, sensitivity, specificity, prediction value positive and prediction value negative, respectively. In conclusion, the proposed method provides a consistent result for segmenting parasite in infected erythrocytes image. This result indicates that this proposed scheme is proper to assist the pathologists in detecting Plasmodium parasites.","PeriodicalId":6598,"journal":{"name":"2019 11th International Conference on Information Technology and Electrical Engineering (ICITEE)","volume":"59 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90625216","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 : 2019-10-01DOI: 10.1109/ICITEED.2019.8929963
Wanthanee Rathasamuth, Kitsuchart Pasupa
Single nucleotide polymorphism (SNP) is a genetic trait responsible for the differences in the characteristics of individuals of a living species. Machine learning has been brought in to classify swine breed according to their SNPs. However, since the number of samples (number of pigs sampled) is usually much smaller than the number of features (SNPs) to classify, there may occur an overfitting problem. Therefore, some feature selection techniques were applied to the entire SNPs to reduce them to a much smaller number of most significant SNPs to be used in the classification. In this study, we used information gain in combination with binary flower pollination algorithm for feature selection as well as a cut-off-point-finding threshold for specifying a 0 or 1 value for a position in the solution vector and a GA bit-flip mutation operator. We called it Modified-BFPA. The classifier was SVM. Evaluated against a few other feature selection techniques, our combination of techniques was, at the very least, competitive to those. It selected only 1.76 % of most significant SNPs from the entire set of 10,210 SNPs. The SNPs that it selected provided 95.12 % classification accuracy. Moreover, it was fast: an average of 1.60 iterations in combination with SVM to find a set of best SNPs that provided the highest classification accuracy.
{"title":"A Modified Binary Flower Pollination Algorithm: A Fast and Effective Combination of Feature Selection Techniques for SNP Classification","authors":"Wanthanee Rathasamuth, Kitsuchart Pasupa","doi":"10.1109/ICITEED.2019.8929963","DOIUrl":"https://doi.org/10.1109/ICITEED.2019.8929963","url":null,"abstract":"Single nucleotide polymorphism (SNP) is a genetic trait responsible for the differences in the characteristics of individuals of a living species. Machine learning has been brought in to classify swine breed according to their SNPs. However, since the number of samples (number of pigs sampled) is usually much smaller than the number of features (SNPs) to classify, there may occur an overfitting problem. Therefore, some feature selection techniques were applied to the entire SNPs to reduce them to a much smaller number of most significant SNPs to be used in the classification. In this study, we used information gain in combination with binary flower pollination algorithm for feature selection as well as a cut-off-point-finding threshold for specifying a 0 or 1 value for a position in the solution vector and a GA bit-flip mutation operator. We called it Modified-BFPA. The classifier was SVM. Evaluated against a few other feature selection techniques, our combination of techniques was, at the very least, competitive to those. It selected only 1.76 % of most significant SNPs from the entire set of 10,210 SNPs. The SNPs that it selected provided 95.12 % classification accuracy. Moreover, it was fast: an average of 1.60 iterations in combination with SVM to find a set of best SNPs that provided the highest classification accuracy.","PeriodicalId":6598,"journal":{"name":"2019 11th International Conference on Information Technology and Electrical Engineering (ICITEE)","volume":"71 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79052468","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 : 2019-10-01DOI: 10.1109/ICITEED.2019.8929935
Shidiq Al Hakim, D. I. Sensuse
Development of applications in knowledge mapping requires a system architecture approach to be able to explain conceptual models that define structure, behaviour and view of the system. There are many system architectures that are made for knowledge mapping but are very diverse, and there are no general guidelines that help a scholar used as references in making them. Therefore this research proposes a universal architectural system that can be used to create a knowledge mapping system. With the literature study method and content analysis on the system architecture used in previous studies, this study proposes four layers which can generally conduct in creating architectural systems, namely: Acquisition layer, database layer, knowledge mapping process layer and user interface layer.
{"title":"Toward a Common System Architecture for Knowledge Mapping","authors":"Shidiq Al Hakim, D. I. Sensuse","doi":"10.1109/ICITEED.2019.8929935","DOIUrl":"https://doi.org/10.1109/ICITEED.2019.8929935","url":null,"abstract":"Development of applications in knowledge mapping requires a system architecture approach to be able to explain conceptual models that define structure, behaviour and view of the system. There are many system architectures that are made for knowledge mapping but are very diverse, and there are no general guidelines that help a scholar used as references in making them. Therefore this research proposes a universal architectural system that can be used to create a knowledge mapping system. With the literature study method and content analysis on the system architecture used in previous studies, this study proposes four layers which can generally conduct in creating architectural systems, namely: Acquisition layer, database layer, knowledge mapping process layer and user interface layer.","PeriodicalId":6598,"journal":{"name":"2019 11th International Conference on Information Technology and Electrical Engineering (ICITEE)","volume":"61 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76943604","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：