Planquarium is a context-aware indoor kitchen garden system, where a user can grow fresh plants and vegetables without prior knowledge. Further, the Planquarium will take care of the plants that are inside it using a Rule-Based Context-Aware environment, that is capable to monitor different aspects of the plant and can provide an ideal environment for the plant inside. Different plants have different requirements therefore, we have integrated profiling systems so that a person can select a plant and the Planquarium will adjust itself accordingly. Initially, we have deployed temperature, humidity, moisture, water and light(Artificial sunlight full spectrum) sensors to monitor the plants. In the future, we can further add soil quality monitors to optimum growth. Planquarium is suitable for congested smart cities, smart homes and for people who care for organic food.
Authors: ijaz uddin (University of Nottingham), altaf uddin (AWKUM), rahat khan (UET peshawar), Rashid Naseem (CUSIT), Arshad Ahmad (University of Sawabi),
Hide Authors & Abstract

The popularity of applying Articial Intelligence to perform prediction, automation tasks has become one of the most conspicuous trends in computer science. However, existent researches in this eld are rarely meet the real-time criteria because of heavy computational tasks. In this work, we proposed to deploy machine learning algorithms which are classication techniques on FPGA platform. The first prototype version based on the architecture with Verilog-HDL is trained and tested by Decision Tree and Neural Network model built based on NSL-KDD dataset. Our experiments with NetFPGA-10G platforms show that the Neural Network core can detect attacks at 1.78 Gbps and up to 9.86 Gbps with packets size from 64B to 1500B, which thoroughly faster the Geforce GTX 850M GPU and i5 8th generation CPU 11x and 83x times respectively. The Neural Network classifier system can function at 104.091 MHz and achieve the accuracy at 87.3%.
Authors: Duc-Minh Ngo (Ho Chi Minh City University of Technology), Binh Tran-Thanh (Ho Chi Minh City University of Technology (HCMUT)), Truong Dang (Ho Chi Minh City University of Technology), Tuan Tran (Ho Chi Minh City University of Technology), Thinh Tran Ngoc (Ho Chi Minh City University of Technology), Cuong Pham-Quoc (Ho Chi Minh City University of Technology (HCMUT) - Vietnam),
Hide Authors & Abstract

Nowadays, the computing is becoming faster and faster to support very other scientific areas. Internet of Thing (IoT) is taking much advantage from it. At the beginning of IoT, the static things joined in IoT such as: cameras, sensors, and vending machines. Due to the progress of computing science, IoT is expanding on mobile things such as cars, patients, cellphones and other mobile things for traffic controlling, health care, or getting information. The network of mobile things is called as Internet of Mobile Things (IoMT). There are some problems to be solved in IoMT as: Security and Privacy, Mobile Data Collection and Analysis. The data collected from the mobile things can help to improve the security and privacy better, or using for special purposes. To get the data of mobile things, moved from one cluster to another one, we need an algorithm to solve following things: mobility of mobile nodes, and changing in number of the mobile nodes. The pi-calculus is one solution for this problem. Pi-calculus is introduced by Milner as a formal language for modeling and verifying system requirements. In this paper, a survey is performed on pi-calculus for IoMT, and other related calculi.
Authors: Phan Cong Vinh (Nguyen Tat Thanh University (NTTU)), Vu Tuan Anh (IUH, Vietnam), Pham Quoc Cuong (HCMUT, Vietnam),
Hide Authors & Abstract

In data visualization, graphs representing multivariable data on multidimensional coordinates shaped cubes enable human to understand better the significance of data. There are various types of cubes for representing different datasets. The paper aims at classifying kinds of cubes to enable human to design cubes representing multivariable datasets. Mathematically, the functional relations among five groups of variables result in the way to structure cubes. The paper classifies cubes as three kinds by the characteristics of datasets, including non-space, 2D-space, and 3D-space multidimensional cubes. The non-space multidimensional cubes are applied for non-space multivariable datasets with variables of objects, attributes, and time. The 2D-space multidimensional cubes are applied for the datasets of movers or objects located on ground at time units. The 3D-space multidimensional cubes are applied for the datasets of flyers or objects positioned in elevated space at time units. The correlation in space and/or time shown on the cubes enables human to discover new valuable information.
Authors: Phuoc Tran (Open University, Hochiminh City, Vietnam), Hong Nguyen (Rubber Industrial College, 1428 Phuriengdo, Dongxoai, Binhphuoc Province, Vietnam), Truong Le (Open University, Hochiminh City, Vietnam), Dang Pham (Nguyen Tat Thanh University),
Hide Authors & Abstract

Flooding is a natural risk, large floods have occurred almost every year. These are major issues that researchers are interested and to identify flooded areas or as-sess the risk of flooding, the researchers using image LiDAR or image RADAR to flood mapping, flood risk management, observation and change detection in floodable area. However, flood modeling or flood assessment don’t solve the problem of flood risks. Therefore, in this paper we propose a new approach of processing methodology based on time series analysis that enables predicting of the floodable areas in the Mekong Delta using new satellite images such as Lansat 7 ETM+, Landsat 8 OLI and sentinel-2 MSI.
Authors: Toan Huynh Phung (CTU, Vietnam), Hiep Huynh (CTU, Vietnam), Tran Loi (FSoft Cantho, Vietnam), Simona Niculescu (Université de Bretagne Occidentale - UBO),
Hide Authors & Abstract

Systems need to be formally verified to ensure that their claimed properties hold at all times of system operation. Deterministic Finite State Machines (FSM) are widely used as model checkers to verify system properties. However, for context-aware systems that have regular inputs and contextual inputs, FSM models become more complex and less intuitive, and do not precisely represent the system behavior. In this paper we use simple examples to introduce the declarative reasoning framework Contelog, a theoretically and practically well grounded work in progress, as a complementary approach that can be used to represent, reason, verify data-centric and contextual properties of context-aware systems.
Authors: Ammar Alsaig (Concordia University), Vangalur Alagar (Concordia University), Shiri Nematollaah (Concordia University),
Hide Authors & Abstract

Hazard map is a major part in the architecture of a Smart City which gives citizens an overview of a particular hazard in different areas within the city. In this paper, we research and implement an IoT-based Air-Pollution Hazard Map System for Ho Chi Minh City. The system consists of Sensor and Gateway nodes, Server, and maps. The collected sensor values include temperature, dust, CO, and CO2. Data collected by Sensor node is transmitted to Gateway and then to Server. Server saves the received data to database and queries according to request from users. A web application have been built to display the data and give users an overview of air pollution state around them.
Authors: Anh Nguyen (Ho Chi Minh City University of Technology), Ri Le (Ho Chi Minh City University of Technology), Loc Nguyen (Ho Chi Minh City University of Technology), Cuong Pham-Quoc (Ho Chi Minh City University of Technology (HCMUT) - Vietnam),
Hide Authors & Abstract

The efficiency of a graph visualizing multivariate data is not only subjectively evaluated by human visual perception but also objectively estimated by visual features of graph. For a designed graph, it is necessary to improve visual features to enable human to extract better information from data. Integrating retinal variables into graph is an approach to increasing visual features of graph. In this study, the constituents of graph are grouped into planar mark classes by qualitative and quantitative characteristics. The retinal variables are studied and structured to integrate into planar mark classes. A process of five steps is proposed to increase visual features by integrating retinal variables into graph. The process is illustrated with two case studies, increasing visual features of bus space-time map with qualitative planar mark classes and increasing visual features of the graph representing the data of hand-foot-mouth epidemic in Binhduong with qualitative and quantitative planar mark classes.
Authors: Phuoc Tran (Open University, Hochiminh City, Vietnam), Hong Nguyen (Rubber Industrial College, 1428 Phuriengdo, Dongxoai, Binhphuoc Province, Vietnam), Ngoc Cam Huynh (Vovankiet High School, Kiengiang, Vietnam), My Thuan Pham (Open University, Hochiminh City, Vietnam), Van Anh Tran (College of Economics, Hochiminh City, Vietnam), Dang Pham (Nguyen Tat Thanh University), Lieu Le (Thu Dau Mot University),
Hide Authors & Abstract

Recommender systems are information filtering software which is capable of resolving the recent issue of internet’s information overload. The recommender system generate the recommendation more suitably based on the data gathered either implicitly like user profile, click information, web log history or explicitly like ratings (scale 1-5), likes, dislikes, feedbacks. The most important challenge to the recommender system is the growing number of online users making it a high dimensional data which leads to the data sparsity problem where the accuracy of recommendation depends on the availability of the data. In this paper, a new approach called formal concept analysis is employed to handle the high dimensional data and a FCA-based recommender algorithm, User-based concept clustering recommendation algorithm (UBCCRA) is proposed. The UBCCRA out performs by accurately generating the recommendation for the group-based users called cluster users. The experimental result is shown to prove the cluster recommendation with good result.
Authors: Chemmalar Selvi G (VIT Vellore), Lakshmi Priya G.G. (VIT Vellore), Rose Joseph (Christ Academy Institute For Advanced Studies),
Hide Authors & Abstract

During the recent years, many different methods of using fuzzy time series for forecasting have been published. However, computation in the linguistic environment one term has two parallel semantics, one represented by fuzzy sets (computation-semantics) it human-imposed and the rest (context-semantic) is due to the context of the problem. If the latter semantics is not paid attention, despite the computation accomplished high level of exactly but it has been distorted about semantics. That means the result does not suitable the context of the problem. Hedge Algebras, an algebraic Approach to domains of linguistic variables, unifying the above two semantics of each term, is the basis of convenient calculation in the language environment and does not distort the semantics of terms. A new approach is proposed through a semantic-based algorithm, focus on two key steps: partitioning the universe of discourse of time series into a collection of intervals and mining fuzzy relationships from fuzzy time series, that outperforms accuracy and friendliness in computing. The experimental results, forecasting enrollments at the University of Alabama and forecasting TAIEX Index, demonstrate that the proposed method significantly outperforms the published ones about accurate level, the ease and friendliness on computing
Authors: Yen Phạm Thế, Loc Vu (Giadinh University, Ho Chi Minh City, Vietnam), Dung Quach (Van Hien University,Ho Chi Minh City, Vietnam),
Hide Authors & Abstract

The objective of this research is to propose a deep learning based-prediction model for pineapple sweetness. In this research, we use a Convolutional Neural Network (CNN) to predict sweetness of pineapples from images. The dataset contains 3,780 pineapple images for training and 1,080 pineapple images for testing. Based on the CNN designed it is found that the best image size is 250 × 250 pixels resized to 25 × 25 pixels. The classification accuracy of training and testing are 72.38% and 78.50%, respectively. In addition, the mean square error values for training and testing are 0.1362 and 0.1156, respectively.
Authors: Adisak Sangsongfa,
Hide Authors & Abstract

It is considered an extended notion of the commutativity of the encryption. Using the computational difficulty of the hidden discrete logarithm problem a new method and post-quantum probabilistic algorithm for commutative encryption are proposed. The finite non-commutative associative algebra containing a large set of the global left-sided unites is used as the algebraic carrier of the proposed method and probabilistic commutative cipher. The latter is secure to the know-plaintext attack and, therefore, efficient to implement on its base a post-quantum no-key encryption protocol. Main properties of the algebraic carrier, which are used in the commutative encryption method, are described.
Authors: Nguyen Minh (Academy of Cryptography Techniques), Moldovyan Dmitriy (Laboratory of Cybersecurity and Post-quantum Cryptosystems St. Petersburg Institute for Informatics and Automation of the Russian Academy of Sciences (SPIIRAS)), Moldovyan Alexandr (Laboratory of Cybersecurity and Post-quantum Cryptosystems St. Petersburg Institute for Informatics and Automation of the Russian Academy of Sciences (SPIIRAS)), Phieu Han (Academy of Cryptography Techniques),
Hide Authors & Abstract

Business cards have been widely used to greet business professionals and exchange contact information. However, the current paper-based way to manage business cards impedes their effective usage, leading to a need for digitalising and extracting business card information. This paper aims to design a business card reader (BCR) application for Android devices. Based on Google vision library, the application digitalises and extracts business card information. We evaluate the application on a dataset of 170 business cards. The results show that the application can digitalise business cards and extract contact information with 88.4% of accuracy. We then further conduct a comparative analysis of our application and other commercial BCR applications. Based on the results, the paper suggests several recommendations for future research.
Authors: Nguyen Hoang Thuan (Can Tho University of Technology), Dinh Thanh Nhan (Can Tho University of Technology), Lam Thanh Toan (Can Tho University of Technology), Nguyen Xuan Ha Giang (Can Tho University of Technology), Quoc Bao Truong (Can Tho University),
Hide Authors & Abstract

Water is essential for human life and socio-economic development. Water pollution is a concern for all mankind. In Vietnam, due to the development of factories and factories, water pollution has become more severe, including the Dong Nai River. In this article, the author uses the concentration of TSS at the Tan Hiep control station in Dong Nai River, using the Kriging interpolation method to find the appropriate model and give the results of water pollution prediction. Dong Nai river area over time with high reliability. TSS data were monitored continuously for three months (from the beginning of February 2018 to the end of April 2018), the predicted results using Kriging interpolation with high accuracy with regression coefficient equal to 1,005, the coefficient is 0.859 (the best value is 1), the forecast error is 2.258, the standard error is 0.044. It shows that using the Kriging interpolation method is an effective and suitable solution in mathematical problems with time information.
Authors: Cong Nhut Nguyen (Auther),
Hide Authors & Abstract

In this paper, we study fuzzy graph represents by using model theory. We use hedge algebra and linguistic variables for modeling and aggregating two graphs. We prove theorem of limiting in models state space. We also figure out preserved property of aggregation operator.
Authors: Nguyen Van Han (College of Science, Hue University), Nguyen Cong Hao (Hue University), Phan Cong Vinh (Nguyen Tat Thanh University (NTTU)),
Hide Authors & Abstract

Geostatistics is briefly concerned with estimation and prediction for spatially con-tinuous phenomena, using data measured at a finite number of spatial locations to estimate values of interest at unmeasured locations. In practice, the costs of in-stalling new observational stations to observe metropolitan water pollution sources, as DO (Dissolvent Oxygen), COD (Chemical Oxygen Demand) and BOD (Biochemical oxygen Demand) concentrations are economically high. In this study, spatial analysis of water pollution of 32 stations monitored during 3 years was carried out. Geostatistics which has been introduced as a management and decision tool by many researchers has been applied to reveal the spatial structure of water pollution fluctuation. In this article, I use the recorded DO con-centrations (is the amount of dissolvent oxygen in water required for the respira-tion of aquatic organisms) at several observational stations on the rivers in Ho Chi Minh City (HCMC), employ the Kriging interpolation method to find suita-ble models, then predict DO concentrations at some unmeasured stations in the city. Our key contribution is finding good statistical models by several criteria, then fitting those models with high precision. From the data set, I found the best forecast model with the smallest forecast error to predict DO concentration on rivers in Ho Chi Minh City. From there we propose to the authorities to improve areas where DO concentrations exceed permissible levels.
Authors: Cong Nhut Nguyen (Auther),
Hide Authors & Abstract

The evolutions in programming from machine language to these days programming software have made it easy, to some extent, to develop software but it is not as easy as programming in natural language. In order to transfer natural language text to any programming language code, it felt necessary to first transform natural language text into pseudo code algorithm then with the help of right API library, such algorithms can be transform into any programming language code. The main aim of this research work is to produce pseudo code from text however this work is very loosely bound to natural language processing. Main components of this proposed work is text analyser that utilizes language tools (spelling check, grammar check) to remove type errors and then eliminate different ambiguities. For this step of ambiguity removal, an adaptive solution is proposed that learning from manual assistance. Once the text is cleared, pattern matching techniques is applied to it and later on parsed into a pseudo code. The concept model is tested with user scenario approach and also practically implemented by developing a prototype. This model is examined using 100 example of different categories and achieved 73 % accuracy. In this era of computing, transformation of natural language text to programming language code is at very early stage thus the proposed model is small contribution toward making such big system.
Authors: Altaf uddin altaf (AWKUM, KP, Pakistan),
Hide Authors & Abstract

The assessment of students' learning ability for career guidance in the future is a huge challenge. The development stage of students' learning ability is considered from the sixth grade to the ninth grade. Student's transcripts from grade 6 to grade 9 are used to assess students' learning abilities. A transcript comparison of grades 6 through 9 is essential for each parent and analyst from there they can guide their children to comprehensive development of knowledge. The objective of this paper is to visually analyze student data using visual analysis approach, proposes a visual analysis system for data discovery with many variables (VAS), a visual data analysis model, visual data analysis criteria, visual data variables, multidimensional cube representing student data, and some visual data analysis questions based on visual graphs related to Junior High School students (JHSSs). Visual analysis of student data helps parents or analysts observe and extract useful information that they interact visual on visual graphs by asking themselves or answering the visual data analysis questions themselves when observing visual graphs by the retina to guide their children to choose the right knowledge chain and future jobs. Visual graphs represent the correlation between subjects and especially the comparison of a subject in the academic years together to help parents and analysts see clearly the trend of the development of students' learning abilities by visual data analysis model.
Authors: Dang Pham (Faculty of Information Technology, Nguyen Tat Thanh University, Hochiminh City, Vietnam), Phuoc Tran (Ho Chi Minh City Open University, Hochiminh City, Vietnam),
Hide Authors & Abstract

The Mekong Delta has made great progress in rice production over the past ten years. Intensive cultivation with multi-cropping brings many benefits to farmers as well as the food export industry. However, this is also an oppor-tunity for raising epidemic outbreak, Brown Plant-hoppers can directly dam-age by sucking the rice’s vitality, and they can cause the wilting and com-plete drying of rice plants, a noncontagious disease known as “Hopper-burn”. In this article, we propose the CDNN model for insect classification based on Neural Network and Deep Learning approach. First, insect images are collected and extracted features based on Dense Scale-Invariant Feature Transform. Then, Bag of Features is used for image representation as feature vectors. Lastly, these feature vectors are trained and classified using CDNN model based on Deep Neural Network. The approach is demonstrated with experiments, and measured by a large amount of Brown Plant-hoppers and Ladybugs samples.
Authors: HIEP HUYNH (Can Tho University, Vietnam), DUY LAM (Mekong University, Vietnam), TU HO (Can Tho University, Vietnam), DIEM LE (Can Tho University, Vietnam), LY LE (Can Tho University, Vietnam),
Hide Authors & Abstract

The relevance of the research is caused by the need of application of speech recognition technology for language teaching. The speech recognition is one of the most important tasks of the signal processing and pattern recognition fields. The speech recognition technology allows computers to understand human speech and it plays very important role in people’s lives. This technology can be used to help people in a variety way such as controlling smart homes and devices; using robots to perform job interviews; converting audio into text, etc. But there are not many applications of speech recognition technology in education, especially in English teaching. The main aim of the research is to propose an algorithm in which speech recognition technology is used English language teaching. Objects of researches are speech recognition technologies and frameworks, English spoken sounds system. Research results. The authors have proposed an algorithm based on speech recognition framework for English pronunciation learning. This proposed algorithm can be applied to another speech recognition framework and different languages. Besides the authors also demonstrated how to use the proposed algorithm for development English pronunciation practicing system based on iOS mobile app platform. The system also allows language learners can practice English pronunciation anywhere and anytime without any purchase.
Authors: Ngoc Hoang Phan (Ba Ria-Vung Tau University), Thi Thu Trang Bui (Ba Ria-Vung Tau University),
Hide Authors & Abstract

Text classification is an essential component in a variety of applications of natural language processing. While the deep learning-based approach is becoming more popular, using vectors of word as an input for the models has proved to be a good way for the machine to learn the relation between words in a document. This paper proposes a solution for the text classification using hybrid deep learning approaches. Every existing deep learning approach has its own advantages and the hybrid deep learning model we are introducing is the combination of the superior features of CNN and LSTM models. The proposed models CNN-LSTM, LSTM-CNN show enhanced accuracy over another approach.
Authors: Hung Bui Thanh (Data Analytics & Artificial Intelligence Laboratory, Thu Dau Mot University),
Hide Authors & Abstract