Vol. 19 No. 4 (2021)

COVID-19 (Coronavirus disease) is an infectious disease caused by a novel strain of Coronavirus that causes a severe acute respiratory syndrome, known as SARS-CoV-2.COVID-19 has posed a serious global public health emergency. Within a short periodof time, the spreadof the virushasbeen recordedin 213 countries. The zoonotic transmission from the bat (bat-to-human) is suspected as the route of disease origin. Furthermore, the emerging public health crisis raises many urgent questions regarding SARS-CoV-2 and animal infection. Several animal models have been reported as candidates for evaluating the efficacy and safety of antiviral drugs or experimental vaccine trials against SARS-CoV-2. Therefore, this review brieflydiscusses COVID-19/SARS-CoV-2 with a particular focus on animal rolesand their susceptibility to the virus.

Food contaminated with heavy metals causes serious consequences for human health, it is always the prime concern of any food safety control systems, even required through costly processes from sample collecting and evaluating contaminated components in the food samples. This study aimed to use ultrasound coupling with deep networks to assess the presence of heavy metals in sweet potatoes, while ultrasound is safe. To classify the acquired ultrasound data sets, we used deep networks that presently become a powerful tool and attract many researchers in order to recognize the data associating with the presence of lead sulfate in samples of sweet potatoes. For the 31 ultrasonic data sets of sweet potato samples acquired, the application of Neuron Network (NN) and Deep Boltzmann machine (DBM) as our target deep networks yielded the results showing that the accuracies of the NN was 62% for training set and 55% for testing set, and of DBM was 68% for training set and 65% for testing set, respectively.

Speech recognition has been attracting many researchers in the field of artificial intelligence recently. For example, the problem of implementing a program for robots to recognize human speech, thereby robots can understand, learn and talk with human. In this study, 37 students from Vietnam National University of Agriculture were involved to acquire speech data of 29 letters in Vietnamese alphabet. The data were preprocessed to extract featured voice chunks for the classification. We then used the deep Boltzmann machine (DBM) as a deep network with stacked hidden layers. To evaluate the proposed method, we compared the learning performance of DBM to a neural network (NN) with the same network structure configuration. The results showed that DBM performed better with accuracies of 68% on the training dataset and 51% on the test dataset, while the respective figures for NN were 61% of training and 48%.

The paper presents a spatio-temporal data forecasting approach using Linear Regression (LR) in two steps called two-step LR. In the first step, all features were divided into subgroups and Linear Regressions was utilized to obtain a regression value for each feature subgroup. In the second step, Linear Regressions was applied again to these regression values to generate the final regression value. The approach using two-step LR had state-of-the-art performance for a wind speed forecasting problem. Wind speed forecasting would be useful for the integration of wind energy into the power grid because wind power generated by wind turbines has an intimate relationship with wind speed and unpredictability and variability of wind speed is one of the fundamental difficulties of this integration system.

Time series with missing values occur in almost areas of applied science. Ignoring missing values can lead to a reduction of system performance and unreliable results, especially in case of large missing values. Therefore, handling missing data is an important task to effectively perform further purposes such as classification, data analysis, etc. This article aims first to introduce approaches for dealing with missing data. Next a framework is built to fill the incomplete data in univariate time series and then to compare the performance of various imputation methods. Four indices are used to evaluate the ability of imputation methods on 3 different real-time data series. Through experimental results, the DTWBI and eDTWBI methods achieve better results with data having seasonality component and without trend factor, while na.interp is more superior as the data have both seasonality and trend components.

Practically, a decision-maker (DM) often faces the problem of choosing the best optimal from the alternatives based on given criteria, which is called the multiple criteria decision-making problem (MCDM). Many models have been studied and used to solve MCDM problems. However, it is wrong to declare that there is a certain method that is the best one. The main objective of this studywas to analyze and evaluate some methods of solving multi-criteria decision-making problems such as the SAW (Simple Additive Weighting), and MOORA (multi-objective optimization based on ratio analysis) method, and apply them to solve some problems in agriculture. To do that, we analyzed the SAW and MOORA methods and proposed improvement directions in the data normalization step and used fuzzy entropy to calculate weights for the model's criteria. After that, the improved models used to solve some problems in agriculture, such as the problem of choosing a mushroom cultivation formula, the problem of choosing a fertilizer for cashew trees. The results obtained after improvement were considered to be better.

The frequent occurrence of poultry diseases, such as bird flu, not only causes huge economic losses to farmers but also poses serious threats to human health. Providing diverse and detailed images of some common diseases in chickens is essential in poultry breeding to reduce economic losses while limiting disease outbreaks on a large scale and mitigating the source of infection from avian to humans. Therefore, in this study, we developed a website of images fully and in detail for some common chicken diseases so that the farmers can easily access and automatically look up information about the considered diseases. The images of the diseased chickenscollected from avian experts were synthesized, standardized and organized into a database of chicken disease images. This is the first database of web-based chicken disease imagesthat replaces the existing paper-based data.

In this study, we compared the performances of two typical frameworks in the server-centric java web framework group, ZK and JSF, towards a usage option. Using a performance test method based on the response time and memory usage criteria, studying simulations, parameter configurations on Jmeter, VisualVM and Java virtual machines were made. The results revealed objective reflection on the performance of the two applications built by each framework. Through the performance test results, we chose ZK, which had a better performance framework than JSF. ZK framework was the applied in practice, building applications that have been put into use at Vietnam National University of Agriculture.

Content-based image retrieval has received great attention in recent years because this method overcomes the disadvantages of the text-based image retrieval that is not affected by the lack of or wrong of the text attached to the image. In addition, deep learning methods such as convolutional neural networks have demonstrated their ability to process large-sized data, especially computer vision and image processing. The aims of this study was develop a content-based image retrieval program and method to reduce image query time using the convolutional neural network (CNN). Also, we combined CNN with a binary hashing method to improve image retrieval time. The experimental results on CIFAR-10 and MNIST data sets showed that combining CNN with the binary hashing method for content-based image retrieval achieved an accuracy of approximately 89% on CIFAR-10,98% on MNIST and significantly improved retrieval time.

The present study aimed to develop a document sharing and management application program for IT majors. Due to students' need to search the documents for learning and researching and to ensure accurate resources, time- and cost-saving, we collected teaching materials from the teachers, then standardized and organized data storage and management uniformly on an online software system to facilitate the search and sharing of documents. The system was built on the PHP technology platform with searching functions which are very flexible and including many different criteria. The diverse shared learning resources were classified into different categories, so it was very convenient for searching. Besides, the document sharing application was not only a centralized and unified document sharing channel but also adhered to the curriculum exclusively for students of the Faculty of Information Technology - Vietnam National University of Agriculture. The program would provide a connection between lecturers and students, brought practical benefits and initially created trust and motivation for students in the process of learning and research.

This study was aimed to provide data on gastrointestinal tract morphology and the relationship of Clark indexand gender, season and sampling site variables of the Linecheek tank goby Glossogobius sparsipapillus. Fish specimens were sampled monthly using trawling nets from September 2019 to August 2020, at three sites in Bac Lieu and Ca Mau provinces. After catching, the samples were determined sex, total length and weight, gastrointestinal tract morphology, gut length and Clark index.The analysis results of 729 specimens showed that this species had a long and pointed snout, flattened head, the mouth was wide and oblique, the lower jaw protruded, and the incision extended to the front eye. There were many rows of teeth on each jaw. It was a carnivorous fish as its intestine length was shorter than the total length of fish (RLG ranged from 0.18to0.83(0.44 ± 0.00)). This species had a high energy storage capacity due to the high value of the Clark index. The Clark of this species varied with season, sampling site, month and season × sampling site, but not gender, gender × sampling site and gender × season. The results have contributed to understanding in future aquaculture of this species.

This research aimedat assessing consumers’ preferences in softshell turtle food consumption and recommending solutionsto meet consumer’s demand. Results on survey of 307 consumers from 40 restaurants on consumer’s characteristics, their references in restaurantsand dish selections and their satisfaction with restaurants’ services indicated that:Softshell turtle food consumers hadmoderate and stableincome;Turtle food was used in the cases of meeting, celebrations, normally at night time;6-8 consumers came at a time by car;Majority of them used turtle food for 3-4 weeks each time;More than 88% chose green turtle with an average of 450-500 gam /person;Preferred dishes includedsoftshell turtle cooked with banana and tofu, roastedwith salt, cooked with medical herbs, steamed with lotus or salt;About38% consumerswere still unsatisfied with services provided by restaurants. The paper also recomends Solutions are therefore reccommended concentrating on middleaged consumer group, both men and women, well-organizing restaurant facilities, car parks, good human resource allocation for table and reception services, development of supply chains in softshell turtles, especially green ones, upgrading knowledge and skills of workers involved in food processing and, right price policies,ensuring sanitation and food... to fully meet consumer preferences.