International Journal of Reasoning-based Intelligent Systems (33 papers in press)
Multimedia-Aided English Online Translation Platform based on Bayesian Theorem
by Xinfei Wang
Abstract: In order to overcome the problems of the traditional online English translation platform, such as low translation efficiency, poor translation accuracy and small translation database capacity, a multimedia aided online English translation platform based on bayesian theorem is designed. The translation platform consists of display layer, permission control layer, logic control layer and data processing layer. This paper introduces bayes' theorem and calculates the probability of translation from English to Chinese. In the design of query module, the translation of search words and thesaurus is selected based on bayes' theorem, and the retrieval method is optimized. Sqlite management system is used to manage the vocabulary data in the vocabulary, so as to complete the design of multimedia assisted English online translation platform. Experimental results show that the translation accuracy of the platform designed in this paper fluctuates in the range of 86-95, and the translation time is always lower than 0.4s, indicating that the platform not only has high translation efficiency and accuracy, but also can complete the translation of large volume data.
Keywords: machine vision technology; multimedia; English; translation platform; platform construction;.
A Neuron based Active Queue Management Scheme for Internet Congestion Control
by Sukant Bisoy
Abstract: To deal with nonlinear and complex problems of Internet congestion control, an intelligent scheme is required, which can learn the traffic pattern of the network. In order to achieve quality of service (QoS) provisioning of real time service, active queue management (AQM) scheme in Internet router needs to stabilize the queue length. In this paper, we design a robust AQM scheme called neuron based AQM (N-AQM) to efficiently control the complex network congestion problem and achieve QoS. In N-AQM, a neural network is used to predict the future value of current queue length and estimate the differential queue length error and use it to define the packet drop probability. The parameters of neurons can be adjusted dynamically depending on the time-varying environment to stabilize the queue length and achieve robustness against different network parameters. Our simulation result demonstrates that N-AQM is stable, robust and outperforms other AQM schemes.
Keywords: AQM; Neuron; QoS; Stability; Robustness.
Fast Recognition Method Of Text Features In Dynamic Video Images
by Guotao Zhao, Jie Ding
Abstract: In order to improve the text recognition ability of dynamic video image, a fast text feature recognition method based on block area contour detection is proposed. The fuzzy video feature analysis method is used to collect the text features of the dynamic video image, reorganize the frame structure, and extract the edge features of the text features of the image. The collected text features are fused into multi-dimensional information, and the edge contour feature information of dynamic video image is matched by block matching method. In this paper, a packet fusion model of Chinese text feature in dynamic video image is established, and the corner detection of text feature is carried out by surf algorithm to realize the fast recognition of Chinese text feature in dynamic video image. Simulation results show that the maximum SNR of the image text feature recognition output is 44.8 dB, and the recognition accuracy is high.
Keywords: dynamic video; image; text feature; fast recognition.
Classification and Prediction of Alzheimer
by Monika Jyotiyana, Nishtha Kesswani
Abstract: As the lifestyle standards of the population increases day by day the risk factor of developing the Alzheimer\'s is increasing and the demands of accurate diagnosis of the disease are also on high demand. This change in population is due to the change in lifestyle and busy schedule. Due to which person is not able to meet their daily requirement for getting fit and in older age maximum population strikes with Alzheimer Disease. In this paper, we have classified the patients suffering from Alzheimer using the NACC dataset with the help of random forest, support vector machine, k- nearest neighbor, linear discriminate analysis and, neural networks. We have used the multilayer perceptron for classification of MRI data and the result obtained showed that the multilayer perceptrons are the most efficient approach for classification with 94% accuracy.
Keywords: Alzheimer\'s disease; Linear Discriminate Analysis; Logistic regression; Random forest; Support vector machine; Classification and prediction.
Towards an interface for translating Natural Language questions to SQL: A conceptual framework from a systematic review
by Karam Ahkouk, Mustapha Machkour
Abstract: Querying Relational Databases using human language is a complex process that must take into consideration the lack of knowledge of SQL by the end users that interact and use the content of these databases systems. Many studies published in the last decade are based on limited, unbalanced or even small datasets, thing that is by far not realistic to train and evaluate a model. The SQL structure is rich and the scale of complexity is very large. Moreover, the difference between the natural language, which is a language of communication, and the SQL language that contains details of implementation specific to relational databases, adds more complications to the problem. In this paper, we present a full systematic review of the existing approaches in the task of text to SQL and we introduce the details of our suggested architecture as well as our new approach to tackle the problem of nested queries.
Keywords: Natural language; machine translation; deep learning; relational databases.
Cryptographic Algorithm for Protection of Communication in Drones Control
by Ivan Ivanov, Krume Andreev, Stella Vetova, Rumen Arnaudov
Abstract: The paper below presents an optimized cryptographic algorithm, designed to protect communication in the management of drones. In this communication the need for the data to be protected as much as possible and (in case) stored reliably rises. The use of microcontrollers requires an optimized algorithm in terms of number and type of operations, due to lower computing power and a small amount of memory. The algorithm is built on DES algorithm, according to the Feistel scheme. It is a 64-bits symmetric block cryptographic algorithm, using a 512-bits cryptographic key.
Keywords: Cryptography; scheme Feistel; cryptographic algorithm; drones; data protection.
Optimization of Multi Branched Horizontal Well Parameters Based on Improved Particle Swarm Optimization Algorithm
by Haiying Ding, Qiang Chen
Abstract: In order to optimize the parameters of multiple branched horizontal well parameters, the improved particle swarm optimization algorithm is put forward through combining the traditional particle swarm algorithm and shuffled frog leaping algorithm, which is applied in solving the optimal model. The optimal model is constructed using the net present value as objective function, and main parameters of multiple horizontal well as optimal variables. Optimal simulation analysis is carried out, and results show that the improved particle swarm optimization algorithm can obtain the optimal effect, which can improve the economical benefit of petroleum industry.
Keywords: Multiple branched horizontal well; Improved particle swarm algorithm; Optimization.
Special Issue on: CIM-19 Advances in Machine Learning and Intelligent Systems - Challenges and Solutions
Energy Efficient Task Scheduling using Adaptive PSO for Cloud Computing
by Rama Rani, Ritu Garg
Abstract: Cloud computing is an important research domain where all computational resources are networked globally and shared to users easily. The cloud service provider (CSP) wants an eco-friendly solution to resolve these issues. To enhance the performance of cloud computing resources, task scheduling is of prime concern. Further, the growth of cloud computing resources leads to a large amount of energy consumption and carbon footprints. Thus, this paper aims to reduce the makespan along with energy consumption for independent tasks. For this purpose, we proposed energy-efficient adaptive particle swarm optimization (EE-APSO) algorithm for independent tasks scheduling decision. Each particle represents a potential solution, and small position value (SPV) rule is used to change the continuous particle position vector to a discrete particle position vector. PSO is made adaptive by varying acceleration coefficients and inertia weight. We also introduced mutation operation to avoid the PSO algorithm getting stuck in local minima and explore the whole search space efficiently. Result analysis demonstrated that our proposed algorithm EE-APSO using SPV rule gives better results than min-min, max-min and genetic algorithm (GA) in terms of makespan and energy consumption.
Keywords: Cloud Computing; Independent task scheduling; Particle Swarm Optimization; Energy Consumption; Makespan.
NITCO: An Intelligent Agent Technique for Optimizing of Resource Utilization in Cloud
by HARVINDER SINGH CHAHAL, ANSHU BHASIN, PARAG RAVIKANT KAVERI
Abstract: Efficient task scheduling is significant to meet the quality of service (QoS) requirements in cloud computing. Cloud is a large pool of virtual access resources to perform thousands of computational and storage operations. Task Scheduling is an NP-hard problem, unsuitable matching leads to performance degradation and violation of service level agreement (SLA). The growing complexity of cloud services needs an extension of existing scheduling algorithms. In this paper, the scheduling problem has been explored based on growing application trends. Cloud dynamic resource provisioning can satisfy users requirements if execution of tasks performed: identifying of task requirements, workflow of application scheduling using a sufficient amount of resources. In this research work, we present an intelligent agent technique for optimizing resource utilization named NITCO. NITCO considers the above mentioned challenge, identification of task requirements and configuration of resource. The performance of proposed NITCO has been evaluated on simulated cloud environmen2016t, and comparison of results show that NITCO performed better in terms of execution cost, execution time, VM utilization and SLA violation while it delivers quality of service.
Keywords: Cloud Computing; Scheduling; Utilization; Energy-consumption; SLA.
Computing Semantic Relatedness by Latent Semantic Analysis and Fuzzy Formal Concept Analysis
by Shivani Jain, Seeja K.R, Rajni Jindal
Abstract: Measuring semantic similarity/semantic relatedness is an important task in Computational linguistic, Natural language processing and Ontology Creation. In this paper, a new hybrid method using LSA and FFCA is proposed for computing the semantic-relatedness. Latent semantic analysis (LSA) is used to extract the attributes of the concepts and these attributes are further mapped to FFCA to compute semantic relatedness. The latent semantic analysis is used for finding the neighboring words or attributes and their correlation value. The concepts and their attributes are mapped to FCA table and then to FFCA table by using the correlation value as membership. A Fuzzy similarity measure is then used to compute the semantic relatedness between these concepts/words. The proposed method is evaluated on word similarity bench mark datasetWS-353 and found an accuracy of 0.85.
Keywords: Semantic Relatedness; Semantic Similarity; LSA; Fuzzy formal concept analysis; fuzzy set similarity measure; Semantic Association;.
ADAPTIVE EDGE-BASED BI-CUBIC IMAGE INTERPOLATION
by C. John Moses, Selvathi D
Abstract: Image interpolation is a technique of creating new pixels by using old pixels. Nowadays image interpolation systems are widely used in many digital signal processing applications like reconstructing medical images and increasing the resolution of satellite and multimedia images. Bi-cubic interpolation is one of traditional and high-performance scheme as compared with other conventional interpolations like nearest neighbour and bilinear. However, the traditional interpolation produces image artifacts like jagging and blurring. To avoid these kinds of drawbacks, several adaptive bi-cubic schemes are introduced in the past decade. This work presents an edge-based adaptive bi-cubic image interpolation using clamp filter and sigmoidal edge detection technique. The clamp filter avoids aliasing artifacts and it smooths edge information by performing low-pass filtering. The experimental result shows that the proposed edge-based bi-cubic outperforms other related bi-cubic image interpolation schemes.
Keywords: Up-scaling; resolution; PSNR; clamp filter; convolution; SSIM; multimedia; bilinear; polynomial; convolution.
Deep Convolutional Neural Network based Diabetic Eye Disease Detection and Classification using Thermal Images
by Selvathi D, Suganya K, Menaka M, Venkatraman B
Abstract: Infrared thermography which is non-contact, non-invasive technique widely accepted as a medical diagnostic tool. Thermal images are processed for abnormality detection and quantification. It has been used in the diagnosis of Dry eye, Meibomian gland dysfunction, Thyroid eye disease and Glaucoma. Diabetic Eye Disease (DED) detection using thermal images is absolutely new attempt. The early detection of the occurrence of DED can be very helpful for clinical treatment. In this paper, we are attempting towards finding an automatic way to classify diabetic eye diseases in thermal images using a deep learning based Convolutional Neural Network (CNN) methodology. The sensitivity of 92.30%, specificity of 98.46%, and accuracy of 95.38% on testing dataset with reference to experts ground truth results are obtained. The results attained evidently exhibit that the thermal imaging is promising modality and proposed deep learning method is capable for automatic diagnosis of diabetic eye disease classification.
Keywords: Convolutional Neural Network; Diabetic Eye Disease; Thermal Imaging; Classification; Ground Truth; Accuracy; Diagnosis; RGB; Testing Dataset; Deep Learning.
Demographical Gender Prediction of Twitter Users using Big Data Analytics: An Application of Decision Marketing
by Sudipta Roy, Bhavya Patel, Debnath Bhattacharyya, Kushal Dhayal, Tai-Hoon Kim, Mamta Mittal
Abstract: The use and influence of digital media, particularly social media, have grown in every sphere of life. One of the trendiest social sites is Twitter. Twitter often contains conversation in non-standard language, and thus, it is difficult to analyze in real-time using conventional language processing. Twitter does not accumulate user gender information as do other popular social media platforms. Thus, demographic feature prediction and additional informative content are important for advertising, custom-made marketing and authorized investigation from the social medium. In this study, proposed statistical representation with real-time analysis using big-data technologies is able to predict the gender of Twitter users. Data cleaning, processing, and storage are performed by the big-data technology Apache Hive. Gender prediction is performed using the naive Bayes classifier to address systemic issues, and Apache Hive is used to solve data storage and big-data processing issues. Authors have considered the tweets-only scenario, the other scenario that was used predicts gender by combining the user tweets and the user profile description. To maintain the stability of the amount of training instances used per estimation, we initiate a balanced class formulation using the polynomial Naive Bayes. Another systemic and previously existing problem of features that was assumed to be independent is solved by the proposed method. The proposed customized method is a speedy, easy-to-implement with pre-processing, close to state-of-the-art document text categorization method using big-data technologies. The proposed statistical method produces higher accuracy in gender classification using tweets only and tweets with description compared with other gold-standard methods.
Keywords: Twitter; Naïve Bayes; Gender Classification; Apache Hive; Perceptron; Logistic Regression;.
Deep Learning based Detection and Prediction of Trending Topics from Streaming Data
by AJEET R.A.M. PATHAK, Manjusha Pandey, Siddharth Rautaray
Abstract: Detecting and predicting trending topics from steaming social data has always been the point of active research area in business and research firms to take quick decisions, change marketing strategies and set new goals. Topic modeling is one of the excellent methods to analyse the contents from large collection of documents in an unsupervised manner and it is a popular method used in natural language processing, information retrieval, text processing and many other research domains. In this paper, deep learning based topic modeling technique has been proposed to detect and predict the trending topics from streaming data. The online version of latent semantic analysis with regularization constraints has been designed using long short-term memory network. Specifically, a problem of detecting the topics from streaming media is handled as the minimization of quadratic loss function constrained by ?1 and ?2 regularization. The online learning mechanism supports scalable topic modeling. For topic prediction, sequence-to-sequence long short-term memory network has been designed. Experimentally, significant results have been achieved in terms of query retrieval performance and topic relevance metrics for topic detection on our published dataset. For topic prediction, the results obtained in terms of root mean squared error are also significant.
Keywords: Deep learning; topic detection; topic prediction; social media data.
Special Issue on: ICCD-2018 Human-Computer Interaction
Generalized Linear Orthomorphisms
by Haiqing HAN, Siru Zhu, Yanqing Dai, Qili MAO, Qin Li, Kang SHI
Abstract: In this scientific research paper, the authors have generalized the concept with regard to orthomorphic permutations(called orthomorphisms) over the Galois field. Meanwhile we have gain the enumeration formula of the total generalised linear orthomorphic permutations over the Galois field, which possesses an arbitrary prime number as the characteristic of the prime subfield. So, the local creating algorithm with regard to partial generalised linear orthomorphic permutations over the Galois general fieldis realized. Comparatively speaking, the innovativeness and originality enumeration formula with regard to linear orthomorphisms over a Galois field with characteristic 2 is a special case to contain in our novel fruits over the general field. What is more, the generalised linear orthomorphic permutations have been thoroughly discussed and generated far
Keywords: P-permutation; Block Cipher; the Branch Number; Generalized Linear Orthomorphism.
Multi-Agent-Based Distributed Text Information Filtering Method
by Wuxue Jiang
Abstract: In order to improve the filtration efficiency and precision, and reduce the occupation of network resources in distributed text information filtering system, a kind of Multi-Agent-based text filtering method was designed. Directed by multi-Agent theory and technology, the system structure and working mechanism of distributed text information filtering are presented, which makes detailed design for scheduling responding agent and learning agent. The load balance was implemented by dynamic range adaptive load migration (DRALM). The experiment shows that this filtering method, boasting higher filtering performance, not only has higher filter precision, but processes tasks in many machines effectively balancing computing load.
Keywords: Multi-Agent System; Text Information Filtering; Distributed System; Open Computing Model; Dynamic Range Adaptive Strategy; Daemon.
Modified Jaya algorithm with chaos
by Mingjing Pei, Shuhao Yu, Maosheng Fu, Xukun Zuo
Abstract: Jaya algorithm is a recently developed optimization algorithm, which is a new optimization algorithm designed to solve optimization related problems, it has two random parameters in equations. In the study of this paper, we will introduce chaos into Jaya so as to increase non-repeatability and ergodicity for global optimization. Here, four different chaotic maps are utilized to control random parameters in Jaya. The results show that some chaotic maps can outperform the random parameters in the high dimensional function and the result of the two-dimensional function is almost the same.
Keywords: Jaya algorithm; Chaos; Global optimization.
Hedging Strategy for Commodity Futures Based on SVM-KNN
by Mei Sun, Rongpu Chen, Yulian Wen, Peiyao Nie
Abstract: In view of the problem of excessive exposure in the field of quantitative investment in commodity futures and policy failure in the low volatility market environment, a new quantitative investment strategy using SVM-KNN combined classifier to hedge multi-factor futures is proposed and applied to the management of quantitative fund. The quantitative investment strategy can not only reduce the overall systemic risk of the investment portfolio, but also adapt to the long-term environment of the commodity futures market. The retest data and the results of real trading show that the SVM-KNN based hedging strategy of commodity futures is significantly higher than the traditional CTA trend tracking strategy in the annual rate of return and the SHARP ratio, and the retracting of the cross period is greatly reduced.
Keywords: Quantitative Investment; Commodity Futures; Multifactor Hedging; Support Vector rn Machine; K-nearest Neighbors.
Special Issue on: ICICT2019 Emerging Technologies for the Internet of Things
Estimating Equations under IPW Imputation of Missing Data
by Hao WU, CuiCui LI, Chen Cheng
Abstract: The IPW imputation method is first applied t to compensate for nonresponse. And then, the empirical likelihood (EL) inference is made for estimation equation parameters. It is a nice result obtained in this paper that the limiting distributions of the EL statistics are 2-type distributions under the IPW imputation. Compared with the usual methods, the proposed method is easier to complement and more efficient.
Keywords: Empirical likelihood; equation estimation; missing data imputation; IPW (inverse probability weighted).
Design and Realization of Vehicle Security and Protection System Based on Multi-task Polling Processing
by Hai-ye Qiao
Abstract: As the increasing number of vehicles and more important role of the car in the daily life, the issue of security and protection of the vehicle has attracted more attention. Nowadays, the functions like location, monitor, anti-theft or the medium, low level vehicles are not popularized. Current existed systems have the problems of incomplete functions, unfriendly- user interface which cannot well fulfill the needs from users. This essay introduces the design of vehicle security and protection system based on C/S architecture. This system could realize the functions of remote monitoring, in-car phone call, fatigue driving monitoring, remote security, alarm call, ensuring the vehicle is under surveillance 24 hours a day, so to enhance the safety factor. This system is proved by the test for its stable running, uninterrupted transmission, user-friend interface and strong operability.
Keywords: Multi-task,Multithreading ,In-car Terminal; GPS; GPRS; C/S.
Research on key indicators and regional comparison of green data center
by Mei Zhang, Fei Feng, Zhi-long Zhang, Jing-hua Wen
Abstract: With the rapid development of information technology and the increasing number of information, green data centers are growing in size, and it is used more and more widely, so green index system of the green data center is the key problem to be solved urgently. This paper studied on the basic index system of green data center, with the help of PUE, CEI and TCO index analysis, it is made in depth analysis of the factors affecting the green index of the green data center, then using factor analysis method from the annual average temperature, annual precipitation, air quality index, seismic belt, information transmission enterprise fixed assets, the number of Internet users, power generation, the whole society and the Internet penetration rate of nine aspects, it is built the mathematical model for evaluating the operational advantages of green data center, finally it is finished that quantitative analysis and evaluation, and according to the calculation of the green data center development area comparison results.
Keywords: Green Data Center; Basic Index System; Green Index System; Regional Comparison; Advantage Evaluation Coefficient.
New Type NP-CSMA of Adaptive Multi-priority Control WSN Protocol Analysis
by Zejun Han, Hongwei Ding, Kun Yue, Liyong Bao, Zhijun Yang
Abstract: Wireless Sensor Networks (WSN) uses a large number of cheap micro sensor nodes, which can be used to perceive and deal with the information that the transmission network covers the object in the geographic area, so it is very popular among people. P-CSMA access to WSN can control the system and reduce energy consumption, but the overall effect is general and the practicability is common. In this paper, a three-clock non-persistent P-CSMA (NP-CSMA) based on variable collision length is proposed to improve system throughput. By dividing the successful packet transmission time (1+a), the collision packet transmission time (b+a), the free packet sending time a, the average cycle method is used to find out the important parameters of the system, such as throughput, collision rate and idle rate, striving to minimize the system energy consumption, effectively extending the life cycle of nodes and reducing costs. Finally, simulation experiments are validated the validity of the agreement, comparing with other protocols to verify the superiority of this agreement.
Keywords: WSN; collision length variable; three-clock; non-persistent P-CSMA; throughput; average cycle method.
Special Issue on: Intelligent Information Technologies and Agriculture
Wireless sensor networks for smart agriculture
by Volodymyr Romanov, Igor Galelyuka, Oleksandr Voronenko
Abstract: Digital agriculture requires development and implementation of new information and communication technologies, including wireless ones for increasing efficiency. The authors developed wireless sensor networks, and wireless technology for express-diagnostics of plant state in agriculture, created the appropriate hardware, software and methodological support. The developed networks now are implemented and under field tests both in open agricultural lands and in smart greenhouses. In this paper the design and application of developed wireless sensor network and some results of calculations and modelling are given
Keywords: wireless sensor network; smart agriculture; Internet of Things.
Strategies based on IoT for supporting the decision making in Agriculture: A Systematic Literature Mapping
by Mario Diván, María Laura Sánchez-Reynoso
Abstract: Smart Agriculture has taken a relevant role due to the importance of food production in contrast to the global population growth. In this sense, alternatives for improving the efficiency and efficacy related to the production systems are welcomed. The Internet-of-Thing (IoT) has allowed increasing the coverage in the field through the use of open-hardware technology at affordable prices. In this way, real-time data collecting has become a feasible alternative for smart agriculture, incorporating different kinds of strategies oriented to real-time decision making with actionable recommendations. As the main contribution, this work implements a systematic mapping study (SMS) of the literature to identify different strategies and approaching oriented to support the data-driven decision making in agriculture using IoT devices. SMS is applied to the Scopus database, with a special focus on those data coming from IoT devices, which eventually could be complemented through big data repositories. 74 documents are retained based on the defined filters in relation to the subject. Also, a scoring model is defined and implemented with the aim of establishing an order for documents, which get a balance between the content precision, the number of citations and the publishing year. The Top-Ten documents are contrasted, while the rest of the documents are synthesized, highlighting their contributions and opportunities to improve in an ordered way based on the scoring model. The documented main subjects related to real-time decision making based on IoT devices are represented by precision agriculture, cultivating monitoring, and irrigation management systems.
Keywords: Strategies; Internet-of-Things; Decision-Making; Agriculture; Systematic Literature Mapping.
The Problems of Consistent Intelligent Real-Time Control of Complex Systems
by Andrey Tyugashev
Abstract: There are a lot of complex technical systems with a growing impact on our life. We can mean Railroad Transportation, Automated Manufactories, Automated Agricultural Systems, and Aerospace Missions, etc. Such systems consist of hundreds of sensors, actuators, aggregates, etc. To complete the specific tasks of the complex, we need to implement a required schedule of logically coordinated actions in real-time mode, even in case of faults of the equipment. The control means should be consistent in various senses defined and discussed in the article. The article further develops and generalizes Real-Time Control Algorithms Logic previously proposed by the author, for considering consistent control in case of limited available resources. The special software tools based on the presented approach are also being described.
Keywords: Intelligent control; Control of Complex Technical Systems; Real-Time Control Algorithms; AI; Program verification; Formal Methods.
Enhanced Interconnected Microgrids for Water-Pumps Networks towards Zero Net Energy (ZNE) Farms
by Ahmed Abdelmaksoud, Hossam A. Gabbar
Abstract: This paper presents design of resilient interconnected microgrids to provide clean and cheap energy to water pumps infrastructure at farms during normal and emergency situations in terms of different renewable resources. Water networks are relied on every day millions of customers utilization and bring billions worth of economic goods to market. However, water network infrastructures are dependent on the electric grid, which is vulnerable to extreme weather, changing supply and demand patterns, and cyber-terrorism. The proposed microgrid includes hybrid energy storage including battery systems, which provides flexible and adaptive energy supply in view of water pumps loads. Proposed microgrid will allow sustainable operation capability with resilient performance on both of economical and technical levels. In addition, the proposed solution, based on Artificial Bee Colony Algorithm (ABC) optimization for Zero Net Energy (ZNE), ensures sustained and high performance energy supply in different operational scenarios. Integrated architecture is proposed to ensure local and supervisory operation and management of the interconnected microgrids, with best use of energy storage systems and energy distributed systems. The proposed designs are modelled and simulated in Simulink, for a proposed system of interconnected microgrids for water-pumps infrastructures, which demonstrates a reduced dependence on the electric grid. The results are both extremely impressive and promising towards a more resilient and stable energy future both for water network and its critical infrastructures.
Keywords: Microgrids; water pumps networks; resilient energy systems; energy storage; renewable resources.
Frost forecast - a practice of machine learning from data
by LIYA DING
Abstract: Abstract: Among the efforts in frost forecast using machine learning techniques, a well adopted method is to first apply time series forecast for the lowest temperature at future time points, such as the next a few days, and then apply predictive model to predict the event of frost at these time points using corresponding temperature forecasted. According to the domain understanding, there exists some cause-effect between environment factors, including temperature and others, and the occurrence of frost in a few hours period. A new modelling concept has been proposed to capture such cause-effect. Preliminary experiments showed encouraging results with a sample of minute-level sensor data collected in Ikuta campus of Meiji University. In this article, as a continuation of the previous work, we shall further discuss methods of modelling, including causal models and associative models, and propose a framework of hybrid system in supporting frost forecast of short-term (e.g. a few hours) as well as that of relatively longer periods (e.g. a few days). More experiments are provided, and the issues of performance evaluation are discussed.
Keywords: Frost forecast; Machine learning; Time series forecasting; Prediction; Cause-effect.
Development of IoT based Smart Agriculture Monitoring System for Red Radish Plants Production
by Ari Aharari, Chunsheng Yang
Abstract: The world population is increasing at a fast rate, and as results need for food is also growing briskly. The traditional method of agriculture is not sufficient enough to cover the needs of the market. On the other hand, the aging of agricultural workers has progressed rapidly, and the successor problem is becoming more serious. Under such circumstances are coming out also new farmers that will help the beginner to agriculture. However, the establishment of farming technology has become a significant management challenge for new farmers. In this paper, we focused on automation in agriculture by applying IoT technologies. The proposed system is utilizing?to monitor the environmental information during the experiment of producing the red radish. The sensor data is analyzed to find the relation between ecological parameters and the growth results. The result of the proposed system was satisfactory as the first step in much deep measure development.
Keywords: Smart agriculture; Internet of Things; Red radish.
Special Issue on: ICEST'19 Reasoning in Engineering Systems Natural and Artificial Intelligence
Low - Cost Energy - Efficient Air Quality Monitoring System Using Sensor Network
by Mare Srbinovska, Aleksandra Krkoleva Mateska, Vesna Andova, Maja Celeska Krsteska, Tomislav Kartalov
Abstract: The air pollution has a significant impact on humans health and global environment. In urban areas the air quality significantly decreases over the past few years. One of the methods for air pollution reduction is by installing a green walls, green roofs or by implementing green buildings in urban areas as plants have capabilities to absorb the particulate matter through their leaves. Urban green spaces especially from trees to green roofs and large natural spaces generally provide significant health benefits for residents and the community. rnThe main goals of this paper is to present system for air quality monitoring using sensor network technology that can be easily deployed in polluted areas; and to examine the influence of the experimental green wall setup to particulate matter concentrations in the air in an urban area in Skopje. Furthermore, the paper presents the preliminary results of the ongoing experiment developed to evaluate the impact of green walls in reduction of air polluting particles. The air quality monitoring system can be easily replicated on other locations in the urban areas of Skopje.rn
Keywords: air quality monitoring system; green walls; sensor network; particulate matter.
A comparative performance analysis of different machine learning techniques for SNR prediction in microcell and picocell wireless environment
by Nikola Sekulovic, Milos Stojanovic, Aleksandra Panajotovic, Milos Bandjur
Abstract: Knowledge of propagation channel conditions enables adaptive data transmission which improves the quality and efficiency of communication system. Wireless channels are characterized by highly dynamic time-varying nature. This means that information regarding propagation channel condition obtained by channel estimation can become outdated because of delay caused by processing and feedback phases. In fast fading environments, prediction of channel based on channel states in previous moments can ensure timely information. In this paper, a comparative performance analysis of an echo state network (ESN), an extreme learning machine (ELM) and least squares support vector machines (LS-SVM) for prediction of wireless channel conditions for single-input single-output (SISO) systems in microcellular and picocellular environments is carried out. Normalized mean squared error (NMSE) and time consumption are used as performance indicators. The experimental results on measured values for signal-to-noise ratio (SNR) show that all models have better and comparable prediction accuracy in microcell environment, while prediction framework based on the ESN outperforms the others in picocell environment.
Keywords: Channel prediction; Echo state network; Extreme learning machines; Least squares support vector machines; Microcellular environment; Picocellular environment.
Use of infrared radiometry in temperature measurement of plant leaf
by Hristo Hristov, Kalin Dimitrov, Stanyo Kolev
Abstract: Through our present work we will show the importance of infrared radiometry in conducting various plant studies. We will look at the factors that affect temperature measurements and their significance. We will monitor how changing the distance between radiometer and object of study affects the heat flow entering the radiometer aperture from an object, and then how changing the distance between them affects the total remaining heat flux entering the radiometer aperture. We will monitor these processes in different temperatures and with different surface areas of the object of study. We will monitor the change of the flux entering from the object of study and the change in the total remaining heat flux entering the radiometer aperture, when the scene is made of insulation material. We will draw conclusions about the significance of the distance between the thermal camera and the object of study.
Keywords: infrared radiometry; infrared thermography; agriculture; solid angle; thermal radiation; emission coefficient; surface temperature.
Special Issue on: CIM-19 Advances in Machine Learning and Intelligent Systems - Challenges and Solutions
A NOVEL APPROACH FOR DYNAMIC INFORMATION INTEGRATION
by Vikash Kumar Garg, Ashish Oberoi, Manish Arora
Abstract: With tremendous growth of data, in recent past, NoSQL databases have emerged and become the popular choices amongst top companies. Scalability can be easily achieved with these databases which is suitable for big companies working with streaming or social data. But the consistency offered by these databases have always remained a concern for companies working with crucial and specially financial transactional data. Moreover the entirely new architecture and new set of query language makes it more expensive and tedious to migrate the existing architecture on this new platform. The proposed architecture in this paper tries to eliminate these business issues by implementing the powers of NoSQL databases using the traditional RDBMS model. With this proposed architecture, the query language remains the same, the ACID properties can be maintained where required and more scalability and reliability can be achieved.
Utilizing Predictive Analytics for Decision Making and improving healthcare services in Public Maternal Healthcare Database
by Shelly Gupta, Shailendra Singh, Parsid Jain
Abstract: Abstract: Background: Predictive analytics is the advanced analytics which is used to make predictions about the unknown future events. In public healthcare datasets, predictive analytics helps to improve the healthcare quality by supporting the healthcare planners in decision making. Hence it is an ongoing research in the field of public healthcare data, especially with the increase in the electronic public healthcare datasets. In this paper the predictive analytics enabled results on public maternal health data (2015-16) of Uttar Pradesh state of India are discussed for enhancing the quality in public maternal healthcare to sustain nation women health. Methodology: In this study a process model based on KDD (Knowledge Discovery in Data) process is presented for predictive analysis in public maternal healthcare data of Uttar Pradesh, state of India. The supervised learning based predictive methods i.e. C4.5, MLP and kNN are used for predictive model building. The dataset is divided in the three major categories i.e. Pregnancy registration and ANC receipt, Deliveries at Home and Medical Facility Availability to achieve the mentioned objectives. Results: The classifier results are compared using the accuracy and error rate matrices of classification. It is found that C4.5 has outperformed over the other two benchmark classifiers. Discussion: In the first category of ANC (Antenatal Care) registration and receipt attributes it is found that the districts with higher percentage of live births rate having weight less than 2.5 kg is an important parameter to be included during NPDs (Non-priority Districts) and PDs (Priority Districts) distribution. Our predictive analysis on deliveries at home category of attributes leads us to know that the PNC (Post Natal Care) checkups within 48 hours percentage are high when deliveries at home are taken under trained SBA (Skilled Birth Attendant). So, this indicates that more effort is needed towards the awareness of the deliveries to be done under trained SBA. In the third category of medical facility availability, it is analyzed that the impact of SCs (Sub-Centres) availability is less to identify priority and non priority districts. Ideally SCs are those care units which have great role in providing the awareness towards health.
Keywords: Predictive Analytics; Public Maternal Health; Machine Learning; ROC (Receiver Operating Characteristic) curve.