The growth of big data market for intelligence-based Internet-of-Things (IoT) users has attracted... more The growth of big data market for intelligence-based Internet-of-Things (IoT) users has attracted both industry and academia. Through using local data from various IoT devices, the service provider can produce valuable information for its users via maching learning (ML) such as centralized learning with a cloud server and local learning with the IoT devices. However, due to privacy leakage risk when the IoT users send the local data to the cloud server and limited computation resources of IoT devices, federated learning (FL) can be the efficient solution to solve the above problems. FL approach is a collaborative ML in which each IoT device can first conduct the individual training process and then share the local model only to the cloud server without data sharing. In this case, this approach can not only improve the training process performance, but also protect data privacy for the IoT users. This research focuses on FL system design with privacy-awareness for IoT users. Particularly, a homomorphic encryption based-encryption method is used to encrypt data from IoT devices during the local training process of the FL as the data privacy protection fom IoT malicious attackers. From this research, we can analyze the model accuracy performance between FL without and with the above encryption method.
Detecting self-care problems is one of important and challenging issues for occupational therapis... more Detecting self-care problems is one of important and challenging issues for occupational therapists, since it requires a complex and time-consuming process. Machine learning algorithms have been recently applied to overcome this issue. In this study, we propose a self-care prediction model called GA-XGBoost, which combines genetic algorithms (GAs) with extreme gradient boosting (XGBoost) for predicting self-care problems of children with disability. Selecting the feature subset affects the model performance; thus, we utilize GA to optimize finding the optimum feature subsets toward improving the model's performance. To validate the effectiveness of GA-XGBoost, we present six experiments: comparing GA-XGBoost with other machine learning models and previous study results, a statistical significant test, impact analysis of feature selection and comparison with other feature selection methods, and sensitivity analysis of GA parameters. During the experiments, we use accuracy, precision, recall, and f1-score to measure the performance of the prediction models. The results show that GA-XGBoost obtains better performance than other prediction models and the previous study results. In addition, we design and develop a web-based self-care prediction to help therapist diagnose the self-care problems of children with disabilities. Therefore, appropriate treatment/therapy could be performed for each child to improve their therapeutic outcome.
Currently, the manufacturing industry is experiencing a data-driven revolution. There are multipl... more Currently, the manufacturing industry is experiencing a data-driven revolution. There are multiple processes in the manufacturing industry and will eventually generate a large amount of data. Collecting, analyzing and storing a large amount of data are one of key elements of the smart manufacturing industry. To ensure that all processes within the manufacturing industry are functioning smoothly, the big data processing is needed. Thus, in this study an open source-based real-time data processing (OSRDP) architecture framework was proposed. OSRDP architecture framework consists of several open sources technologies, including Apache Kafka, Apache Storm and NoSQL MongoDB that are effective and cost efficient for real-time data processing. Several experiments and impact analysis for manufacturing sustainability are provided. The results showed that the proposed system is capable of processing a massive sensor data efficiently when the number of sensors data and devices increases. In addition, the data mining based on Random Forest is presented to predict the quality of products given the sensor data as the input. The Random Forest successfully classifies the defect and non-defect products, and generates high accuracy compared to other data mining algorithms. This study is expected to support the management in their decision-making for product quality inspection and support manufacturing sustainability.
An improved service of carsharing, one-way service enables customers to use the vehicles from one... more An improved service of carsharing, one-way service enables customers to use the vehicles from one station and return to other station. The common issue in one-way carsharing service is that the vehicle stock of each station become imbalance, thus will lead to the less customer satisfaction and less utilization of vehicles. Consequently, the relocation is used by the system to move the appropriate vehicle to a high demand station in order to elevate customer satisfaction. This paper will demonstrate the periodically relocation model for a one-way carsharing system and tested on simulation model. Computational simulation result on commercially operational data of carsharing in South Korea is involved to solve the relocation problem in one-way system. The results we have obtained in this study provide a clear insight into the impact of model on low relocation cost compared to the traditional relocation.
An improved service of carsharing, one-way service enables customers use the vehicles from one st... more An improved service of carsharing, one-way service enables customers use the vehicles from one station and return to other station. The common issue in one-way carsharing service is that the vehicle stock of each station become imbalance, thus will lead to the less customer satisfaction and less utilization of vehicles. Consequently, the relocation is used by the system to move the appropriate vehicle to high demand station in order to elevate customer satisfaction. This paper will demonstrate the Ant Colony Optimization for relocation in one-way carsharing system based on simulation model. Computational simulation result on commercially operational data of carsharing in South Korea is involved to solve relocation problem in one-way system. The result we have obtained in this study provide a clear insight into the impact of Ant Colony Optimization for relocation problem on utilization of system and high customer satisfaction.
2019 5th International Conference on Science and Technology (ICST), Jul 1, 2019
Radio frequency identification (RFID) technology can be utilized to monitor tagged product moveme... more Radio frequency identification (RFID) technology can be utilized to monitor tagged product movements and directions for the purpose of inventory management. It is important for RFID gate to identify the several RFID readings such as movement type and direction as well as the static tags (tags that accidentally read by the reader). In this study, random forest (RF) method is utilized to detect the movement type and direction of RFID passive tags. The input features are derived from received signal strength (RSS) and timestamp of tags. The result showed that machine learning models successfully distinguish direction and movement type of tag. In addition, proposed model based on random forest generated accuracy as much as 98.39% and was significantly superior to the other models considered.
Biocybernetics and Biomedical Engineering, Oct 1, 2020
Predicting future blood glucose (BG) levels for diabetic patients will help them avoid potentiall... more Predicting future blood glucose (BG) levels for diabetic patients will help them avoid potentially critical health issues. We demonstrate the use of machine learning models to predict future blood glucose levels given a history of blood glucose values as the single input parameter. We propose an Artificial Neural Network (ANN) model with time-domain attributes to predict blood glucose levels 15, 30, 45 and 60 min in the future. Initially, the model's features are selected based on the previous 30 min of BG measurements before a trained model is generated for each patient. These features are combined with time-domain attributes to give additional inputs to the proposed ANN. The prediction model was tested on 12 patients with Type 1 diabetes (T1D) and the results were compared with other data-driven models including the Support Vector Regression (SVR), K-Nearest Neighbor (KNN), C4.5 Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost) and eXtreme Gradient Boosting (XGBoost) models. Our results show that the proposed BG prediction model that is based on an ANN outperformed all other models with an average Root Mean Square Error (RMSE) of 2.82, 6.31, 10.65 and 15.33 mg/dL for Prediction Horizons (PHs) of 15, 30, 45 and 60 min, respectively. Our testing showed that combining time-domain attributes into the input data resulted in enhanced performance of majority of prediction models. The implementation of proposed prediction model allows patients to obtain future blood glucose levels, so that the preventive alerts can be generated before critical hypoglycemic/ hyperglycemic events occur.
Lecture notes in electrical engineering, Nov 23, 2016
Food quality and safety has gained main attention, due to increasing health awareness of customer... more Food quality and safety has gained main attention, due to increasing health awareness of customer, improved economic standards and lifestyle of modern societies. Thus, it is important for consumers to purchase good quality products in order to keep the customer satisfaction level. In this study, we propose traceability system for food by monitoring the location as well as temperature and humidity. The RFID technology and wireless sensor network are utilized in this study to perform the experiment. The real testbed implementation has been performed in one of the Korean Kimchi Supply Chain. The result showed that our proposed system gave the benefit to the manager as well as customer by providing real time location as well as temperature-humidity history. It will help manager to optimize the food distribution while for the customer it will increase the satisfaction by maintaining the freshness of product.
Understanding customer shopping behavior in retail store is important to improve the customers’ r... more Understanding customer shopping behavior in retail store is important to improve the customers’ relationship with the retailer, which can help to lift the revenue of the business. However, compared to online store, the customer browsing activities in the retail store is difficult to be analysed. Therefore, in this study the customer shopping behavior analysis (i.e., browsing activity) in retail store by utilizing radio frequency identification (RFID)-enabled shelf and machine learning model is proposed. First, the RFID technology is installed in the store shelf to monitor the movement tagged products. The dataset was gathered from receive signal strength (RSS) of the tags for different customer behavior scenario. The statistical features were extracted from RSS of tags. Finally, machine learning models were utilized to classify different customer shopping activities. The experiment result showed that the proposed model based on Multilayer Perceptron (MLP) outperformed other models by as much as 97.00%, 96.67%, 97.50%, and 96.57% for accuracy, precision, recall, and f-score, respectively. The proposed model can help the managers better understand what products customer interested in, so that can be utilized for product placement, promotion as well as relevant product recommendations to the customers.
Indonesian Journal of Electrical Engineering and Computer Science
Machine learning algorithms can be used to forecast future blood glucose (BG) levels for diabetes... more Machine learning algorithms can be used to forecast future blood glucose (BG) levels for diabetes patients, according to recent studies. In this study, dataset from continuous glucose monitoring (CGM) system was used as the sole input for the machine learning models. To forecast blood glucose levels 15, 30, and 45 minutes in the future, we suggested deep neural network (DNN) and tested it on 7 patients with type 1 diabetes (T1D). The suggested prediction model was evaluated against a variety of machine learning models, such as k-nearest neighbor (KNN), support vector regression (SVR), decision tree (DT), adaptive boosting (AdaBoost), random forest (RF), and eXtreme gradient boosting (XGBoost). The experimental findings demonstrated that the proposed DNN model outperformed all other models, with average root mean square errors (RMSEs) of 17.295, 25.940, and 35.146 mg/dL over prediction horizons (PHs) of 15, 30, and 45 minutes, respectively. Additionally, we have included the suggeste...
Indonesian Journal of Electrical Engineering and Computer Science
Federated learning (FL) has emerged as one of the most effective solutions to deal with the rapid... more Federated learning (FL) has emerged as one of the most effective solutions to deal with the rapid utilization of internet of things (IoT) in big data markets. Through FL, local data at each IoT device can be trained locally without sharing the local data to the cloud server. However, this conventional FL may still suffer from privacy leakage when the local data are trained, and the trained model is shared to the cloud server to update the global prediction model. This paper proposes a FL framework with privacy awareness to protect data including the trained model for IoT devices. First, a data/model encryption method using fully homomorphic encryption is introduced, aiming at protecting the data/model privacy. Then, the FL framework for the IoT with the encryption method leveraging logistic regression approach is discussed. Experimental results using random datasets show that the proposed framework can obtain higher global model accuracy (up to 4.84%) and lower global model loss (up...
Advances in Computer Science and Ubiquitous Computing, 2016
The RFID technology can be used for item tracking and inventory control. However, the problem suc... more The RFID technology can be used for item tracking and inventory control. However, the problem such as miss reading and ghost reading usually occur in RFID implementation and has impact on low accuracy of inventory management system. In this study, the computer vision is used to solve the problem of miss reading and ghost reading in RFID. The RFID and computer vision can act as ears and eyes respectively, thus by combining both technologies; it is expected to increase the accuracy of system. The result of experiment has showed that the combination of RFID and computer vision has increased the system accuracy, as the computer vision can help the RFID system to detect the miss reading and ghost reading.
2019 5th International Conference on Science and Technology (ICST), 2019
Radio frequency identification (RFID) technology can be utilized to monitor tagged product moveme... more Radio frequency identification (RFID) technology can be utilized to monitor tagged product movements and directions for the purpose of inventory management. It is important for RFID gate to identify the several RFID readings such as movement type and direction as well as the static tags (tags that accidentally read by the reader). In this study, random forest (RF) method is utilized to detect the movement type and direction of RFID passive tags. The input features are derived from received signal strength (RSS) and timestamp of tags. The result showed that machine learning models successfully distinguish direction and movement type of tag. In addition, proposed model based on random forest generated accuracy as much as 98.39% and was significantly superior to the other models considered.
The growth of big data market for intelligence-based Internet-of-Things (IoT) users has attracted... more The growth of big data market for intelligence-based Internet-of-Things (IoT) users has attracted both industry and academia. Through using local data from various IoT devices, the service provider can produce valuable information for its users via maching learning (ML) such as centralized learning with a cloud server and local learning with the IoT devices. However, due to privacy leakage risk when the IoT users send the local data to the cloud server and limited computation resources of IoT devices, federated learning (FL) can be the efficient solution to solve the above problems. FL approach is a collaborative ML in which each IoT device can first conduct the individual training process and then share the local model only to the cloud server without data sharing. In this case, this approach can not only improve the training process performance, but also protect data privacy for the IoT users. This research focuses on FL system design with privacy-awareness for IoT users. Particularly, a homomorphic encryption based-encryption method is used to encrypt data from IoT devices during the local training process of the FL as the data privacy protection fom IoT malicious attackers. From this research, we can analyze the model accuracy performance between FL without and with the above encryption method.
Detecting self-care problems is one of important and challenging issues for occupational therapis... more Detecting self-care problems is one of important and challenging issues for occupational therapists, since it requires a complex and time-consuming process. Machine learning algorithms have been recently applied to overcome this issue. In this study, we propose a self-care prediction model called GA-XGBoost, which combines genetic algorithms (GAs) with extreme gradient boosting (XGBoost) for predicting self-care problems of children with disability. Selecting the feature subset affects the model performance; thus, we utilize GA to optimize finding the optimum feature subsets toward improving the model's performance. To validate the effectiveness of GA-XGBoost, we present six experiments: comparing GA-XGBoost with other machine learning models and previous study results, a statistical significant test, impact analysis of feature selection and comparison with other feature selection methods, and sensitivity analysis of GA parameters. During the experiments, we use accuracy, precision, recall, and f1-score to measure the performance of the prediction models. The results show that GA-XGBoost obtains better performance than other prediction models and the previous study results. In addition, we design and develop a web-based self-care prediction to help therapist diagnose the self-care problems of children with disabilities. Therefore, appropriate treatment/therapy could be performed for each child to improve their therapeutic outcome.
Currently, the manufacturing industry is experiencing a data-driven revolution. There are multipl... more Currently, the manufacturing industry is experiencing a data-driven revolution. There are multiple processes in the manufacturing industry and will eventually generate a large amount of data. Collecting, analyzing and storing a large amount of data are one of key elements of the smart manufacturing industry. To ensure that all processes within the manufacturing industry are functioning smoothly, the big data processing is needed. Thus, in this study an open source-based real-time data processing (OSRDP) architecture framework was proposed. OSRDP architecture framework consists of several open sources technologies, including Apache Kafka, Apache Storm and NoSQL MongoDB that are effective and cost efficient for real-time data processing. Several experiments and impact analysis for manufacturing sustainability are provided. The results showed that the proposed system is capable of processing a massive sensor data efficiently when the number of sensors data and devices increases. In addition, the data mining based on Random Forest is presented to predict the quality of products given the sensor data as the input. The Random Forest successfully classifies the defect and non-defect products, and generates high accuracy compared to other data mining algorithms. This study is expected to support the management in their decision-making for product quality inspection and support manufacturing sustainability.
An improved service of carsharing, one-way service enables customers to use the vehicles from one... more An improved service of carsharing, one-way service enables customers to use the vehicles from one station and return to other station. The common issue in one-way carsharing service is that the vehicle stock of each station become imbalance, thus will lead to the less customer satisfaction and less utilization of vehicles. Consequently, the relocation is used by the system to move the appropriate vehicle to a high demand station in order to elevate customer satisfaction. This paper will demonstrate the periodically relocation model for a one-way carsharing system and tested on simulation model. Computational simulation result on commercially operational data of carsharing in South Korea is involved to solve the relocation problem in one-way system. The results we have obtained in this study provide a clear insight into the impact of model on low relocation cost compared to the traditional relocation.
An improved service of carsharing, one-way service enables customers use the vehicles from one st... more An improved service of carsharing, one-way service enables customers use the vehicles from one station and return to other station. The common issue in one-way carsharing service is that the vehicle stock of each station become imbalance, thus will lead to the less customer satisfaction and less utilization of vehicles. Consequently, the relocation is used by the system to move the appropriate vehicle to high demand station in order to elevate customer satisfaction. This paper will demonstrate the Ant Colony Optimization for relocation in one-way carsharing system based on simulation model. Computational simulation result on commercially operational data of carsharing in South Korea is involved to solve relocation problem in one-way system. The result we have obtained in this study provide a clear insight into the impact of Ant Colony Optimization for relocation problem on utilization of system and high customer satisfaction.
2019 5th International Conference on Science and Technology (ICST), Jul 1, 2019
Radio frequency identification (RFID) technology can be utilized to monitor tagged product moveme... more Radio frequency identification (RFID) technology can be utilized to monitor tagged product movements and directions for the purpose of inventory management. It is important for RFID gate to identify the several RFID readings such as movement type and direction as well as the static tags (tags that accidentally read by the reader). In this study, random forest (RF) method is utilized to detect the movement type and direction of RFID passive tags. The input features are derived from received signal strength (RSS) and timestamp of tags. The result showed that machine learning models successfully distinguish direction and movement type of tag. In addition, proposed model based on random forest generated accuracy as much as 98.39% and was significantly superior to the other models considered.
Biocybernetics and Biomedical Engineering, Oct 1, 2020
Predicting future blood glucose (BG) levels for diabetic patients will help them avoid potentiall... more Predicting future blood glucose (BG) levels for diabetic patients will help them avoid potentially critical health issues. We demonstrate the use of machine learning models to predict future blood glucose levels given a history of blood glucose values as the single input parameter. We propose an Artificial Neural Network (ANN) model with time-domain attributes to predict blood glucose levels 15, 30, 45 and 60 min in the future. Initially, the model's features are selected based on the previous 30 min of BG measurements before a trained model is generated for each patient. These features are combined with time-domain attributes to give additional inputs to the proposed ANN. The prediction model was tested on 12 patients with Type 1 diabetes (T1D) and the results were compared with other data-driven models including the Support Vector Regression (SVR), K-Nearest Neighbor (KNN), C4.5 Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost) and eXtreme Gradient Boosting (XGBoost) models. Our results show that the proposed BG prediction model that is based on an ANN outperformed all other models with an average Root Mean Square Error (RMSE) of 2.82, 6.31, 10.65 and 15.33 mg/dL for Prediction Horizons (PHs) of 15, 30, 45 and 60 min, respectively. Our testing showed that combining time-domain attributes into the input data resulted in enhanced performance of majority of prediction models. The implementation of proposed prediction model allows patients to obtain future blood glucose levels, so that the preventive alerts can be generated before critical hypoglycemic/ hyperglycemic events occur.
Lecture notes in electrical engineering, Nov 23, 2016
Food quality and safety has gained main attention, due to increasing health awareness of customer... more Food quality and safety has gained main attention, due to increasing health awareness of customer, improved economic standards and lifestyle of modern societies. Thus, it is important for consumers to purchase good quality products in order to keep the customer satisfaction level. In this study, we propose traceability system for food by monitoring the location as well as temperature and humidity. The RFID technology and wireless sensor network are utilized in this study to perform the experiment. The real testbed implementation has been performed in one of the Korean Kimchi Supply Chain. The result showed that our proposed system gave the benefit to the manager as well as customer by providing real time location as well as temperature-humidity history. It will help manager to optimize the food distribution while for the customer it will increase the satisfaction by maintaining the freshness of product.
Understanding customer shopping behavior in retail store is important to improve the customers’ r... more Understanding customer shopping behavior in retail store is important to improve the customers’ relationship with the retailer, which can help to lift the revenue of the business. However, compared to online store, the customer browsing activities in the retail store is difficult to be analysed. Therefore, in this study the customer shopping behavior analysis (i.e., browsing activity) in retail store by utilizing radio frequency identification (RFID)-enabled shelf and machine learning model is proposed. First, the RFID technology is installed in the store shelf to monitor the movement tagged products. The dataset was gathered from receive signal strength (RSS) of the tags for different customer behavior scenario. The statistical features were extracted from RSS of tags. Finally, machine learning models were utilized to classify different customer shopping activities. The experiment result showed that the proposed model based on Multilayer Perceptron (MLP) outperformed other models by as much as 97.00%, 96.67%, 97.50%, and 96.57% for accuracy, precision, recall, and f-score, respectively. The proposed model can help the managers better understand what products customer interested in, so that can be utilized for product placement, promotion as well as relevant product recommendations to the customers.
Indonesian Journal of Electrical Engineering and Computer Science
Machine learning algorithms can be used to forecast future blood glucose (BG) levels for diabetes... more Machine learning algorithms can be used to forecast future blood glucose (BG) levels for diabetes patients, according to recent studies. In this study, dataset from continuous glucose monitoring (CGM) system was used as the sole input for the machine learning models. To forecast blood glucose levels 15, 30, and 45 minutes in the future, we suggested deep neural network (DNN) and tested it on 7 patients with type 1 diabetes (T1D). The suggested prediction model was evaluated against a variety of machine learning models, such as k-nearest neighbor (KNN), support vector regression (SVR), decision tree (DT), adaptive boosting (AdaBoost), random forest (RF), and eXtreme gradient boosting (XGBoost). The experimental findings demonstrated that the proposed DNN model outperformed all other models, with average root mean square errors (RMSEs) of 17.295, 25.940, and 35.146 mg/dL over prediction horizons (PHs) of 15, 30, and 45 minutes, respectively. Additionally, we have included the suggeste...
Indonesian Journal of Electrical Engineering and Computer Science
Federated learning (FL) has emerged as one of the most effective solutions to deal with the rapid... more Federated learning (FL) has emerged as one of the most effective solutions to deal with the rapid utilization of internet of things (IoT) in big data markets. Through FL, local data at each IoT device can be trained locally without sharing the local data to the cloud server. However, this conventional FL may still suffer from privacy leakage when the local data are trained, and the trained model is shared to the cloud server to update the global prediction model. This paper proposes a FL framework with privacy awareness to protect data including the trained model for IoT devices. First, a data/model encryption method using fully homomorphic encryption is introduced, aiming at protecting the data/model privacy. Then, the FL framework for the IoT with the encryption method leveraging logistic regression approach is discussed. Experimental results using random datasets show that the proposed framework can obtain higher global model accuracy (up to 4.84%) and lower global model loss (up...
Advances in Computer Science and Ubiquitous Computing, 2016
The RFID technology can be used for item tracking and inventory control. However, the problem suc... more The RFID technology can be used for item tracking and inventory control. However, the problem such as miss reading and ghost reading usually occur in RFID implementation and has impact on low accuracy of inventory management system. In this study, the computer vision is used to solve the problem of miss reading and ghost reading in RFID. The RFID and computer vision can act as ears and eyes respectively, thus by combining both technologies; it is expected to increase the accuracy of system. The result of experiment has showed that the combination of RFID and computer vision has increased the system accuracy, as the computer vision can help the RFID system to detect the miss reading and ghost reading.
2019 5th International Conference on Science and Technology (ICST), 2019
Radio frequency identification (RFID) technology can be utilized to monitor tagged product moveme... more Radio frequency identification (RFID) technology can be utilized to monitor tagged product movements and directions for the purpose of inventory management. It is important for RFID gate to identify the several RFID readings such as movement type and direction as well as the static tags (tags that accidentally read by the reader). In this study, random forest (RF) method is utilized to detect the movement type and direction of RFID passive tags. The input features are derived from received signal strength (RSS) and timestamp of tags. The result showed that machine learning models successfully distinguish direction and movement type of tag. In addition, proposed model based on random forest generated accuracy as much as 98.39% and was significantly superior to the other models considered.
Uploads
Papers by Ganjar Alfian