This paper proposes a new approach for academic advising, especially for  undergraduate groups, based on the approach of maximal tolerance class in the set-valued information system to build groups of students with similar learning outcomes and training results. Based on these results, the academic advisors can organize and propose solutions for each group of similar students in an effective manner, and help the students with poor academic performance to find the solutions to overcome their weakness. At the same time, the  academic advisors can also identify the similar groups of good learning outcome students so that the advisors can build groups for studying and doing scientific research to provide opportunities for students getting better in studying.

To protect a computer system from threat infections, an anti-virus system needs to scan for malicious codes which may appear in target systems. In this paper, we present a technique to recognize malicious codes quickly by using hash indexing mechanism in categorical space. First, in training phase, the dataset of malicious codes will be separated into clusters which have the same characteristics. After that, we build a special rule set in a hash indexing format of ordered cluster-buckets. Next, in recognition phase, we extract object?s characteristics and code them into a checksum value by using some popular hash algorithms. Then this value is used as a key to search for the same rule in indexed rule space. Finally, the system returns the scanning process results. We have built this technique for D2 Anti-virus* 2013 running on Windows XP SP3 in a computer with Intel Core 2 Duo E7200 ? 2.53 GHz. Using a dataset of 615,880 malicious signatures, D2 needs only 05 seconds to examine 105,330 MB of 8,696 executable files. Its average scanning speed is about 21,651MBps. The experimental results denote that this is an effective technique to improve the scanning speed of anti-virus systems nowadays.

Building WebGIS system integrated database management provides the Department of Agriculture and Rural Development of Can Tho city data management (or call AGRIWEBGIS_CT) for the following six departments: Irrigation, Fisheries, Animal Health, Rural Development, Plant Protection, and Water Supply. AGRIWEBGIS_CTis built using ASP.NET Framework programming, database management system with new features in SQL SERVER 2008 version has an integrated spatial database and the open source library SharpMap. The system interfaces Vietnamese, many utility and ease of use as thematic maps to view. This system also provides aggregate statistics of each field data and option that allows system administrators to update the data of the departments. Besides, the general information page supports users during access as rural agriculture news.

MODIS (MODerate Resolution Spectroradiometer) Remote Sensing Technology with varying multispectral spatial and temporal resolution and large coverage, which suitable for monitoring the changes of Earth's surface in large scale. Some kind of data from satelite images provides information to monitor vegetation cover such as MOD09Q1 image, with spatial resolution of 250x250m, and at 8-day interval, suitable for the study of crops  and rice cropping calendar. Based on the images properties (Normalized Difference Vegetatione Index-NDVI), the relationship with the change of rice crop status in space and time to help identify the rice cropping calendar and cropping season (case study in An Giang province 2012-2013). Besides, the the proediction of pest occurences could be generated based on the correlation between the rice growing stages and pests occurence. Results show that can use satellite images of MODIS-MOD09Q1 to generate the rice cropping stages and pest occurrens status. It has a close correlation between vegetation indices with different rice growing stages at high accuracy (R2 = 0.83).

In recent years, climate change is one of the environmental problems that needs to be studied in the Mekong Delta of Vietnam, especially those in conjunction with temperature and rainfall. As temperature and rainfall changes directly affect agriculture and aquaculture activities - driving factors of the delta?s development, the raising question is if such changes could be forecasted with acceptable level of uncertainties. This paper presents algorithms and models of adjusting the forecasted rainfall data obtained from climate data of the SEA-START. A comparison of these forecast models is conducted by forecast error analysis. A case study is experimented by using rainfall data in Can Tho city - Vietnam. The results show that the linear regression model has the greatest forecast error while the non-linear forecast models give better results. The diversity of these forecast models can be applied to solve environmental problems in practice.

Today with the development of DNA sequencing technology, we have obtained a large amount of DNA sequences in a short time with low cost. Specially, the next-generation DNA sequecing can generate a huge amount of short DNA sequences, called short reads with length from 30 to 100 bp. The short reads have an error rate between 1% and 2%. Therefore, the error reads must be corrected before being assembled into the complete genome. There are several proposed algorithms for correcting the error reads such as SHREC and SOAP de Novo. However, SHREC needs a long computation time to correct errors while SOAP de Novo requires very high memory usage. In this paper, we present our algorithm (RCorrector) based on the index structure of KMER for detecting and correcting error reads. Compared to the SHREC algorithm, the RCorrector algorithm provides a speed up from 3 to 7 with the same sensitivity and specificity.

Web filtering is used to prevent access to black web pages (web pages have desirable content or images). In this paper, we apply classification method with Support vector machine learning (SVM) to build a web filtering tool that is integrated with 2 filters: text filter – use text classification method and image filter – use image classification method. With two filters, this tool can prevent user access to desirable content web pages or remove desirable images when web page is displayed on the browser.

In this paper, we present a solution for the scheduling problem of pratical courses in Can Tho University using Maximum Flow Network approach. The scheduling problem for practical courses is concerned to assign students to groups/practical rooms under some constraints such as room capacity, teachers? schedule, etc. The problem attempts to optimize the performance criteria and distribute the students fairly to rooms depending on the ratio of room capacity and the number of students enrolled. We model the scheduling problem as a maximum flow problem where each student is modeled as a source node and each room as a sink node. Capacity constraints are represented as maximum capacity of edges connecting source nodes and sink nodes. We have also added constraints on minimum number of students assigned to a room as minimum capacity of edges. The approach is implemented on Google computing platform using Google Apps Script. A case study is experimented by enrollment data of practical courses in Can Tho University. Results show that our solution is tractable.

In recent years cloud computing has emerged as a new development phase of the Internet. Cloud computing, its service models in particular,  allows the use of hardware, software as services, that leads a fundamental change in the application of information technology in practice - shift from investment to the hire of computing resources. However, cloud adoption is still controversy issues surrounding the safety and security of system. This paper provides an overview on safety issues and security of cloud computing, from the perspective of architecture, services as well as main characters of cloud computing.

Recently, multipath transport control protocol (multipath TCP) allows spreading its data packets on several paths simultaneously. Such the multipath transfer can improve TCP throughput, can balance congestion among paths, and can provide native handover in a network. Current coupled multipath congestion control (MPTCP) was designed for back-compatibility with single-path TCP Reno and for general applications. However, data rate of MPTCP has high variance that not suitable for multimedia applications which require smooth data rate. In this paper, we propose an extension algorithm of single path TFRC, named MPTFRC, designed from both the analytical model of TCP Reno at flow level and the technique of flappiness prevention between paths at packet level. The simulation results demonstrate that the proposed MPTFRC algorithm not only satisfies the three design goals of multipath congestion control algorithm but also provides data rate smoother than that of MPTCP while preservingfair sharing to the existing TCP Reno and MPTCP flows.

This paper describes the design, implementation, and deployment of wireless sensor network for precision agriculture at Long An University of Economics and Industry. Precision agriculture is a farming management concept which applies the right amount of input (water, fertilizer, pesticide, etc.) at the right location and at the right time to enhance production and improve quality, while protecting the environment. Wireless sensor network is built for monitoring and controlling environmental parameters, such as the  environmental parameters (air temperature, air humidity, light, insects, weeds, etc.) and other parameters related to soil conditions (soil moisture, pH, etc.). The data are collected, stored and transmitted wirelessly to the farmers to handle, through which they can control and take appropriate actions for their farm to increase production and quality. From the hardware side, the system consists of three components: wireless sensor nodes DHLA-WSN, wireless sensor node manager DHLA-WMN, and a server controller. Software was built at each node to carry out their task. The deployed system is testing on the field, working properly and promising which brings significantly benefits to the agriculture field.

In model checking, the state space explosion prevents verifying the large systems because exhaustive search fails to find errors. The problem becomes much more serious recently when the size and the complexity of the system-under-check are constantly growing up. Abstraction is the most effective method to overcome this problem. It is an approach used to reduce the size (of the state space) of the model in the reality. However, how to use abstraction effectively in reducing model checking efforts is still a question. This work proposes a multiple abstraction refinement technique in symbolic model checking that allows to verify large state space systems. The proposed technique employs the abstraction refinement algorithm CEGAR proposed by Clarke et al. in 2000 combining with the multiple abstraction method of Qian and Nymeyer. Instead of checking on concrete model, an abstract model is verified first. If a counter-example is found, the searching scope will be widen on the next abstract model and so on, even on the original model. This process can be called refinement. The experimental results show that the new approach improves the performance of the model checking process.

Knowledge base is a central part of intelligent system, so knowledge representation plays an important role in constructing these systems. Nowadays, there are many various methods for representing knowledge. In fact, a popular form of knowledge domain is knowledge about relations, however, the current models are not efficient for executing and reasoning on this kind of knowledge domain. Besides that, automatic drawing geometric objects is an important and basic tool in solving geometry problems. In this paper, we present a method for modelizing the relations between geometric objects. Base on this model, we also introduce an application of this drawing method for solving plane geometry problems in middle school program.

Automated Guided Vehicle (AGV) is an intelligent machine that has enough intelligence to determine its motion status according to the environment conditions. TypicalAGV has the ability to operate and move purposely without human intervention. In this paper, we present a simple method based on computer vision techniques which help the vehicle to move inside the lane boundaries. The vision image will be processed to detect the lane boundaries using vector-lane-concept and estimate the road lane curvature. Using this information, we calculate the steering angle which is used to steer the vehicle to move inside the lane boundaries without using any other control technique. Several demonstrations were carried out, using diverse images, to show the result of lane-boundary detection and vehicle navigation. In conclusion, the proposed algorithm can be used to control vehicle to move inside the lane boundaries without human intervention.

A dataset is called imbalance if it has some classes containing more instances than others. In this case, accurately classifying samples in small classes is very difficult. The higher the imbalanced ratio, the more difficult getting a good solution. Cost-sensitive learning is an effective solution for the imbalanced problem. In this paper, we present a decision system with misclassification cost. The system improves the degree of precision in the minor classes which are interested in imbalanced dataset.The system is based on the study of methods of classifying on the imbalanced dataset by cost-sensitive. This system is applied in medical diagnostic. The experimental results show that the accuracy of the diagnostic system is improved.

In this paper, we present how to combine the machine learning algorithms into Pioneer P3-DX Robot that control the Robot's behaviors, movements, and tracking according to the real time. The machine learning algorithm, Cascades of Boosted Classifiers with Haar-Like Feature, can be used to detect and regconize objects. Next, we used Camshift Algorithms (Continiously Adaptive Meanshift) to monitor the object's activities within a flat image. Finally, based on the object's size and position in the flat image, we control the Robot to move and keep track of the object. The experimental results so that the Robot has been operated as the target objectives such as controlling behaviors, moving and keeping track of an object.

To help programmers quickly develop their software systems, research communities have proposed some approaches for automatic search of function APIs, mostly based on keywords used in the API descriptions. However, the accuracy of those approaches is still limited. Formal specification with the rigor of the syntax and semantics can be applied to improve the research results of this direction. In this paper, we firstly present a survey on formal specification techniques for software design. Next, we propose a framework for searching and reusing API functions automatically based on formal specification described by JML. In addition, the framework also supports automatic composition of API functions to meet the requirement demand. Some case studies are also presented to imply the practical potential of our approach.

The congestion problem happens when there is a crowd accessing online has become more frequent. There have been several different solutions to solve this problem, such as upgrading equipments which have capable of faster processing, or using cloud computing with simultaneous virtual server number increasing with the number of online instant crowd. Another aspect of the congestion problem is to  forecast itself for having time to prepare response measures, especially for dealing with DDoS attacks by hackers. Based on the queuing network model for a system serving online crowd, we have modeled congestion  threshold survey. The application presents results of the survey and the predicted congestion thresholds of  online course registration system at Can Tho University. These results allow to propose a new online course registration  system based on cloud computing with reasonable configuration.

Radio-Frequency Identification (RFID) and Near Field Communication (NFC) technologies applied in e-payment systems have been employed in many countries around the world for a variety of areas such as paying fees in public transportations, buying foods at fast food restaurants, at super markets, and buying soft drink at vending machines. In recent years, these technology solutions have been used in some regions in Vietnam through paying bills at parking services, vending machines, school and company canteens? These technologies make our life easier and more convenient for consumers, save people?s times, reduce cash transactions, centrally manage public services, and modernize our daily life. This paper presents the prospects of applying the RFID and NFC technologies to the e-payment methods illustrated through the design of an e-payment system for motorcycle parking services. The results of this work have demonstrated the feasibility of implementing the system?s hardware in Vietnam.

Recommender system can provide suitable items to users by using data about their behavior in the past to predict the future items that users may like. Two successful approaches in recommender system (relying on the collaborative filtering) are the latent factor models which identify potential relationships on both the user and the item; and neighbor models which use similarities between the items or the similarities between the users. In this study, we introduce an approach which is based on the method proposed by Koren (2010) to utilize the advantages of both the aforementioned approaches. Moreover, besides building a web-based movie recommender system, we try to improve the prediction results by adding to the original model several new regularization coefficients for different models? parameters.

In this paper, we introduce extended episode model upgraded from episode pattern in time-related data. Based on this model, we present an algorithm that finds all frequently extended episodes from an input event sequence without rescaning. By application using new characteristics of mined extended episodes, we propose an application in the diabetes data*. Experimental results of this article show that the extended episodes contain useful information for prediction models.

This article focuses on evaluating students' study results based on weighted average combined with ordered weighted averaging by using WOWA operator (Weighted Ordered Weighted Averaging operator). The research also concentrates on the theory and application problems of the OWA operator (Ordered Weighted Averaging operator) and the extended WOWA operator aims to solving some practical issues based on the study results of the undergraduate students. The weights obtained in this research is determined based on the input information (ordered by statistics) for easy observation, and also for emphasizing the importance of considerably high (or low) values when collecting data of students. Together with the known average calculating method, this study also provides a new method to identify the average based on statistics, which is applied to evaluate the students' study results by combining weighted average (widely used) with the ordered weighted average. Early results of this study has proved the importance of applying integrated operators when collecting information from varied sources, criterions and purposes.

To assist the reader faces the information explosion, we built the recommender system applied for a news website automatically (NewsRES). The NewsRES based on the content-based method and collaborative method. The content-based method is used in comparison the content of information or describing news in order to find out the similar news which the users used to be concerned. The CF method passes the tastes of users to take advice or predictions about unknown tastes for other users. The system is applied to 280 students grade 10, 11 at Le Anh Xuan high school for a week.  We gain the results: 30.59% of precision, 94.17% of recall and 45.26% of F-measure.