Channa striata or the striped snakehead fish is one of snakehead fish species which inhabits all types of freshwater bodies distributed across Asian countries. Because this fish is known to have higher albumin fractio...
Channa striata or the striped snakehead fish is one of snakehead fish species which inhabits all types of freshwater bodies distributed across Asian countries. Because this fish is known to have higher albumin fraction (64.61%) of protein and other economic values, domestication, and cultivation of this fish has been done in many Asian countries such as Indonesia, China, Malaysia, Thailand, Bangladesh, and India. Environmental factors such as temperature, water pH, dissolved oxygen, total dissolved solids, and turbidity are important parameters must be considered inbreeding and growing this type of fish. The aim of this paper is to propose an IoT solution to automatically monitor these environmental factors. It is designed with affordable and open-source electrical components to provide a cost-efficient solution for farmers. Five sensors are used to measure each parameter. A web application prototype is also presented as a companion application for the users to get useful information from the IoT device. It is developed using a Pythonframework. By accessing this web application, the users can immediately detect any abnormal conditions of the pond.
Matrix factorization (MF) technique has been widely utilized in recommendation systems due to the precise prediction of users' interests. Prior MF-based methods adapt the overall rating to make the recommendation ...
ISBN:
(数字)9781728128207
ISBN:
(纸本)9781728128214
Matrix factorization (MF) technique has been widely utilized in recommendation systems due to the precise prediction of users' interests. Prior MF-based methods adapt the overall rating to make the recommendation by extracting latent factors from users and items. However, in real applications, people's preferences usually vary with time; the traditional MF-based methods could not properly capture the change of users' interests. In this paper, by incorporating the recurrent neural network (RNN) into MF, we develop a novel recommendation system, M-RNN-F, to effectively describe the preference evolution of users over time. A learning model is proposed to capture the evolution pattern and predict the user preference in the future. The experimental results show that M-RNN-F performs better than other state-of-the-art recommendation algorithms. In addition, we conduct the experiments on real world dataset to demonstrate the practicability.
COVID-19 makes the community must carry out activities such as school, work, and worship at home. However, the long-running activities from home make people experience boredom which can lead to stress. On the other si...
COVID-19 makes the community must carry out activities such as school, work, and worship at home. However, the long-running activities from home make people experience boredom which can lead to stress. On the other side, the entertainment obtained by the public through smartphones by reading articles by their interests can reduce boredom. This paper proposed a rules-based decision support system to help the people to make choices of their activities from home while COVID-19 e rules-based approach for make an application decision support system. Ruled base used in application to selected process through characteristics following the interests of the community. This decision support system is implemented in mobile web applications. The system can display articles based on interests by questions or statements through the front end system. The results showed that most users of the application in a happy condition while working from home.
Artificial Bee Colony (ABC) is one of good heuristic intelligent algorithm to solve optimization problem including clustering. Generally, the heuristic algorithm will take the high computation time to solve optimizati...
详细信息
Understanding the reliability of engineering methods is crucial for its adoption and deployment. This research focuses on the reliability of the Power Spectral Density (PSD) method via the use of the F statistic for d...
Understanding the reliability of engineering methods is crucial for its adoption and deployment. This research focuses on the reliability of the Power Spectral Density (PSD) method via the use of the F statistic for damage detection. To the author best knowledge, the method is rather classic but its realibility has not been discussed in the context of a large data size. Priory, the research anticipates that the accuracy is a function of the damage level. In this study, we evaluate 3500 cases with five levels of structural integrity, namely, healthy condition and damaged conditions with 1%, 5%, 10%, and 20% damage levels. The dataset is established via a numerical analysis of a seven degree-of-freedom system loaded with a concentrated dynamic force with random magnitude. A spring on the system is reduced in its stiffness to simulate damages. Our significant findings are the following: it is challenging for the PSD-based method to differentiate the healthy condition from the damaged conditions when the damage level is small. However, the reliability is high at 95% probability when the structural integrity has dropped by five percent.
Inserting data into digital media tends to change the large size of the media cover file, this is because the inserted pixel has a different size from the insertion value, i.e. with a larger or smaller value, for exam...
Inserting data into digital media tends to change the large size of the media cover file, this is because the inserted pixel has a different size from the insertion value, i.e. with a larger or smaller value, for example a pixel value of 10 is inserted with larger or smaller data from the value 10. So that the pixel value is getting further from the value of 10, this causes the value of Mean Squared Error (MSE) of the media of the cover (cover) to become larger. In this study, the selection of pixels in the image is in accordance with the value of the insertion data as the insertion place with the Genetic algorithm (GA). Insertion of data on selected pixels using the GA algorithm is performed using the Modified Least Significant Bit (MLSB) algorithm at the position of the least significant bits. To find out the reliability of the proposed algorithm, the data insertion experiment was carried out by comparing the conventional LSB algorithm with the proposed algorithm where the results obtained a significant MSE value, i.e. the average MSE value of the LSB algorithm was 20.54 and the proposed algorithm was 7.12.
For damage detection, this research article discusses an easy-to-compute damage index derived from the governing dynamic of the structure that has potential practical application in Structural Health Monitoring (SHM)....
For damage detection, this research article discusses an easy-to-compute damage index derived from the governing dynamic of the structure that has potential practical application in Structural Health Monitoring (SHM). The research uses simplified structural models to explore the sensitivity of the index to damages, to compare the index performance with a traditional but popular damage detection method, and to understand the local/global predictive capability of the index. The research uses two simple models, namely, single- and two-degree-of-freedom systems. The results suggest that the damage index is local, that can only monitor damages occurring near the points of measurements, but it is sensitive to damages, unlike the natural frequency, which is global but less sensitive.
The success of learning in a classroom that uses supporting media demands lecturer activity. The level of activity of lecturers in class and using learning media such as Smart lecturer can be measured using a fuzzy lo...
The success of learning in a classroom that uses supporting media demands lecturer activity. The level of activity of lecturers in class and using learning media such as Smart lecturer can be measured using a fuzzy logic approach. This study aims to measure the level of active lecturers using the approach, namely: Mamdani and Sugeno method. Stages of lecturer activity measurement by forming a fuzzy set, composition rules as many as 24 rules and the defuzzification process using the centroid method that produces the level of activity of each is low, medium and high. Based on 173 lecturer activity data, the results of the Mamdani method indicates low 66%, medium 31%, and high 3%. While the Sugeno method produces a level of activity low 75%, moderate 17%, and high 8%. Therefore, Mamdani method is more suitable for the calculation because the spread of results is relatively evenly distributed at each level.
The purpose of this study is to look for factors, indicators, and build a model of readiness for effective and efficient implementation in order to answer the problems that arise at PT XYZ related to the readiness of ...
The purpose of this study is to look for factors, indicators, and build a model of readiness for effective and efficient implementation in order to answer the problems that arise at PT XYZ related to the readiness of Customer Relationship Management (CRM) implementation. The research method used is the method of collecting data by observing, interviewing, and distributing questionnaires to respondents using the CRM Value Chain theory as a conceptual framework and the research method used in this research is factor analysis to process data obtained from questionnaires. In this study managed to get new factors along with the constituent indicators of these factors as well as an ideal readiness model. It can be concluded that factors related to the readiness of implementing Customer Relationship Management among others are Customer Service Information, Customer Relations Value, Lack of Service and Communication.
Breast cancer is one of the deadliest cancer for female nowadays. Despite of the rapid advancement in medical image analysis with the rise of deep learning, development of breast cancer detection system is limited due...
详细信息
Breast cancer is one of the deadliest cancer for female nowadays. Despite of the rapid advancement in medical image analysis with the rise of deep learning, development of breast cancer detection system is limited due to relatively small size of the publicly available mammogram dataset. In this paper, we discover an effective configuration for transfer learning from Chest X-Ray pre-trained Convolutional Neural Network to overcome the small-size mammogram dataset problem. We found that the best configuration achieve 90.38% validation accuracy for modified.
暂无评论