Showing papers in "DEStech Transactions on Computer Science and Engineering in 2017"

Electricity Consumption Prediction Using XGBoost Based on Discrete Wavelet Transform

Abstract: The purpose of this paper is to predict the daily electricity consumption of the next month. It is considerably important for people to cope with the problem well. Although few articles mentions the topic of electricity consumption prediction, numerous papers include some topic similar to the topic in this paper, such as rainfall forecasting, wind speed prediction and water flow forecasting. Moreover, a number of techniques and algorithms are employed to cope with those issues and achieve outstanding performance. Those techniques and algorithms are considerably remarkable, but the accuracy of them is not excellent enough on the long-term prediction of time series. In this paper, we propose a hybrid model which integrate discrete wavelet transform and XGBoost to forecast the electricity consumption time series data on the long-term prediction, namely DWT-XGBoost. The original time series data can decompose into approximate time series data and detail time series data by the discrete wavelet transform. And those time series data by decomposition are as features input into the prediction model that is XGBoost. Furthermore, the parameters of XGBoost are obtained by a grid search method. The performance of the proposed model in this paper is measured against with other hybrid models such as integrating discrete wavelet transform and support vector regression, integrating discrete wavelet transform and artificial neural networks, and unitary XGBoost. The comparison results show that the DWT-XGBoost outperforms other models and is a novel method on the long-term prediction of time series.

Parameter Tuning in Random Forest Based on Grid Search Method for Gender Classification Based on Voice Frequency

Abstract: Parameter optimization is one of methods to improve accuracy of machine learning algorithms. This study applied the grid search method for tuning parameters in the well-known classification algorithm namely Random Forests. Random Forests was implemented on the voice gender dataset to identify gender based on the human voice’s characteristics. There are two parameters that were tuned to obtain the optimal values. Those parameters are number of variables used in building trees and number of trees that involves in the classifiers. Experimental results on voice gender dataset show that the highest accuracy of Random Forest with parameter tuning is 0.96907 which is higher than the accuracy of the model without parameter tuning (0.9675). The optimal parameter for the best classifier is number of variables is 'sqrt' which is square root of parameters involved in dataset and number of trees is 300. This study shows that the tuning parameter results optimal parameters for developing the best classifier using Random Forests.

Inter-Blockchain Communication

Abstract: With the rapid development of Blockchain technology, researchers have devoted much to it, and many exciting achievements have been made in industry. However, it doesn’t work well with the massive transactions, leading to islands of redundant and inconsistent information arising. Therefore, a mechanism on how to make different Blockchain systems communicate with each other better, and the Inter-Blockchain communication (IBC) will be proposed in this paper. In IBC, the nodes between the Blockchain systems use the efficient routing algorithm to update the routing path dynamically. It will analysis the transaction and pack the transaction into a specific form to adapt to the destination Blockchain system. In this way, IBC can make different Blockchain system communicate without any intermediaries. The experiment results show that the use of IBC does not have much impact on throughput of Blockchain system.

Research on the Innovation of College Students’ Ideological and Political Education in Big Data Era

Abstract: Based on the new situation and requirements of big data era, this paper aims at the relations of schools, teachers and students, revolves the socialist core values of the "Ideological Education lifeline", puts forward the important principles of the innovation of the ideological and political education of college students, thinks of specific methods and measures for further work of the big data era was proposed to ideological and political education.

Three Approaches to the Gyroid Structure Modelling as a Base of Lightweight Component Produced by Additive Technology

Abstract: The article deals with the modelling of gyroid structure as a base of lightweight component produced by additive technology. Gyroid is one type of so called porous structure that can give to the product extraordinary combination properties such are high strength, stiffness along with low weight and good absorption of energy. This is a reason, why they are widespread not only in mechanical, aerospace or automotive industries, but also in biomedicine. In mechanical engineering, it is possible to produce the gyroids by some of modern additive technologies, because the complex surface of this type it is not possible to produce in another way. The basement of product, which is made by some of additive technologies, is virtual 3D model, so it is necessary to know how to generate it. Three approaches to the modelling have been presented in the article, including the description of the advantages or disadvantages of individual method. Well prepared model can be used not only for gyroid structure manufacturing, but also for simulation its behavior in real practice.

Critical Literature Review of Theory of Planned Behavior in the Information Systems Research

Abstract: The Theory of Planned Behavior (TPB) has been widely applied in several disciplines to help better understand individual human behavior. For the past two decades, the TPB theory have been used in information systems research for providing the ability to predict a wide range of individual behaviors as they are exposed to different technologies and process in organizations and society. The TPB predictive power has been useful in better understanding issues driving individual human behavior as they interact with information technologies. Despite TPB’s impact to the information systems discipline, other researchers suggested additional elements to the basic model in order to better understand complex human behavior. In addition the TPB theory has seen addition of constructs to come up with new frameworks and models such as Technology Acceptance Model (TAM). Besides some criticisms of TPB and limitations it is still widely applied in several disciplines which include information systems.

A Novel Kinship Verification Method Based on Deep Transfer Learning and Feature Nonlinear Mapping

Abstract: There are some problems when the discriminative features are used in the traditional kinship verification methods, such as focusing on the local region information, containing a lot of noisy in non-face regions and redundant information in overlapping regions, manual parameters setting and high dimension. To solve the above problems, a novel kinship verification method based on deep transfer learning and feature nonlinear mapping is proposed in this paper. Firstly, a new deep learning model trained on the face recognition dataset is transferred to the kinship datasets to extract high-level feature. Secondly, siamese multi-layer perceptrons and triangular similarity metric learning are combined to reduce the dimensionality of feature vector by nonlinear mapping. Meanwhile it would guarantee a smaller distance between kin pairs while a larger distance between non-kin pairs. Lastly, the cosine similarity of feature vector pairs is computed, and traditional classifier, such as SVM, is used. Experiments on the TSKinFace, KinFace W-I and KinFace W-II datasets indicate the proposed method could achieve better performance than the traditional methods.

Intrusion Detection of the ICS Protocol EtherCAT

Abstract: Control mechanisms like Industrial Controls Systems (ICS) and its subgroup SCADA (Supervisory Control and Data Acquisition) are a prerequisite to automate industrial processes. While protection of ICS on process management level is relatively straightforward—well known office IT security mechanisms can be used—protection on field bus level is harder to achieve as there are real-time and production requirements like 24x7 to consider. One option to improve security on field bus level is to introduce controls that help to detect and to react on attacks. This paper introduces an initial set of intrusion detection mechanisms for the field bus protocol EtherCAT. To this end existing Ethernet attack vectors including packet injection and man-in-the-middle attacks are tested in an EtherCAT environment, where they could interrupt the EtherCAT network and may even cause physical damage. Based on the signatures of such attacks, a preprocessor and new rule options are defined for the open source intrusion detection system Snort demonstrating the general feasibility of intrusion detection on field bus level.

QISLSQb: A Quantum Image Steganography Scheme Based on Least Significant Qubit

Abstract: In this paper, a new quantum image steganography scheme to embed quantum secrete gray image into quantum cover image is proposed. In the proposed scheme, the quantum secret image scrambled utilizing Arnold cat map and embedding the result into quantum cover image using two least significant qubit (LSQb).The extracting process need only stego image to extract the embedded secret image. The simulation results demonstrate that the proposed scheme has good invisibility and high capacity.

Sea Clutter and Target Detection with Deep Neural Networks

Abstract: In this paper, we investigate the feasibility of applying a deep learning approach to sea clutter suppression and target detection in an inhomogeneous oceanic environment. The employed method consists of deep convolutional auto-encoders (DCAEs) to filter sea clutter and a logistic regression classifier to achieve the detection of target. The sea clutter suppression processing using DCAEs automatically removes complex patterns like superimposed clutter from a target, rather than simple patterns like echoes missing at random. Compared with conventional methods for sea clutter suppression, the algorithm does not need to estimate the covariance matrix of clutter so as to have better flexibility. The results show that reliable suppression performance and higher detection accuracy can be achieved from our experiments, whose data include the measured data and simulation data.

EFQM Model and Sustainability of Organizations

Abstract: The longevity of organizations depends on the ability to deliver quality products or services to the customers. Starting with the 1920s the notion of quality in organizations has gained importance and currently millions of companies worldwide apply one model of quality management or the other. European Foundation for Quality Management (EFQM) model is a prime example practiced throughout Europe. With the world resources being ever depleted at fast rate sustainability is also becoming a key issue in today’s business world. Sustainability and corporate social responsibility principles are generally in line with the EFQM model and have many common elements. This paper explains and demonstrates how the EFQM criteria and sustainability principles overlap. Examples will be given from reverse logistics and supply chain practices.

Learning for Tower Detection of Power Line Inspection

Abstract: Power line inspection is very important for electric company to keep good maintenance of power line infrastructure and ensure reliable electric power distribution. Research efforts focus on automating the inspection process by looking for strategies to satisfy all kinds of requirements. Following this direction, this paper proposes a learning approach for tower detecting problem where aggregate channel features are used to train the boost classifier. Adopting the sliding window paradigm, the electric tower can be located very fast. The main advantages of this approach are its efficiency and accuracy for processing huge quantity of image data. Obtaining highly encouraging results shows that it is really a promising technique.

Application of Autoencoder in Depression Diagnosis

Abstract: Major depressive disorder (MDD) is a mental disorder characterized by at least two weeks of low mood which is present across most situations. Diagnosis of MDD using rest-state functional magnetic resonance imaging (fMRI) data faces many challenges due to the high dimensionality, small samples, noisy and individual variability. No method can automatically extract discriminative features from the origin time series in fMRI images for MDD diagnosis. In this study, we proposed a new method for feature extraction and a workflow which can make an automatic feature extraction and classification without a prior knowledge. An autoencoder was used to learn pre-training parameters of a dimensionality reduction process using 3-D convolution network. Through comparison with the other three feature extraction methods, our method achieved the best classification performance. This method can be used not only in MDD diagnosis, but also other similar disorders.

Phishing Website Detection Using C4.5 Decision Tree

Abstract: Among various anti-phishing solutions, Machine Learning techniques are considered to be promising. The purpose of this study was to evaluate the effect of C4.5 decision tree algorithm on phishing website detection. In the experiment, C4.5 had learned two models using two phishing website datasets respectively, which were PWD and PWD2, the latter was obtained through dimensionality reduction on the former. Under 10-fold cross-validation, various metrics indicated that the two models all done very well. The results of the corrected paired t-test under the significance of 0.05 (two tailed), shown that accuracy and recall metrics of the model based on PWD is not statistically significantly different to those of PWD2. According to the better model, the one based on PWD2, which had the lower complexity and the similar performance compared with the one based on PWD, the top 5 key features for classification were obtained.

Fine-Tuning Parameters of Deep Belief Networks Using Artificial Bee Colony Algorithm

Abstract: The Deep Belief network has become a powerful tool in nowadays to large-scale-oriented application, however, there are several parameters need to assign in advances that is key factors of successive application. In this paper, we proposed to address the issue of properly fine-tuning parameters of Deep Belief Networks by means of Artificial Bee Colony (ABC) algorithm. Experimental results show that the proposed ABC algorithm can effectively reconstruct the original binary images.

Research of Test of Shear Resistance of Modified Emulsified Asphalt by Waterborne Epoxy Resin

Abstract: In view of diseases like slippage and surge of pavement structure that result from being lack of interlayer cohesive force performance of asphalt pavement, interlayer shear stress of five kinds of typical pavement structure was calculated by means of finite element software Ansys; the shear fatigue performance and shear strength performance of epoxy emulsified asphalt were tested on the basis of calculation of interlayer maximum shear stress and shear strength. Results indicate that the shear strength increases with the increase of normal stress by approximate linearly trend; the optimum epoxy emulsified asphalt distributing amount is 0.8kg/m2 under the high temperature of 60℃; the shear strength decreases with the increase of temperature, while the distributing amount and normal stress are the same, in fact, it is mainly caused by the decrease of cohesive force; the maximum interlayer shear stress under the loads of 100kN and 200kN are less than allowable shear stress and shear strength, shear fatigue damage and ultimate shear failure under the condition of high temperature and heavy load will not be generated on the asphalt surface course.

Chinese Text Detection and Recognition in Natural Scene Using HOG and SVM

Abstract: It remains an important, yet challenging problem to detect and recognize text from natural scene images. Since earlier this year a number of methods have been proposed for reading text from scene images. But all of them are adapted for English alphabets, and not suitable for Chinese characters, which present unique challenges including highly versatile fonts, complex background and uneven illumination, a huge number of different characters, unconnected strokes within a character. In this paper we design a method tailored for reading Chinese characters from natural scenes. During the phase of character detection, we employ MSER (Maximally Stable Extremal Region) method to extract candidate characters, and then integrate extracted strokes through mathematical morphology computation. Based on the attributes of Chinese characters, we also lay down heuristic rules that distinguish text and non-text to screen the region of a candidate character. In this method, we describe the features of characters with HOG descriptor and accurately use SVM (Support Vector Machine) according to classification. And the positive region is the region covering the text. At the stage of text recognition, we use KNN (K-Nearest Neighbors) as the classifier. We test this method on 400 natural scene images containing Chinese characters collected from different sources. And the results have successfully validated the efficacy of our approach.

Speech Emotion Recognition Using Convolutional- Recurrent Neural Networks with Attention Model

Abstract: Speech Emotion Recognition (SER) plays an important role in human-computer interface and assistant technologies. In this paper, a new method is proposed using distributed Convolution Neural Networks (CNN) to automatically learn affect-salient features from raw spectral information, and then applying Bidirectional Recurrent Neural Network (BRNN) to obtain the temporal information from the output of CNN. In the end, an Attention Mechanism is implemented on the output sequence of the BRNN to focus on target emotion-pertinent parts of an utterance. This attention mechanism not only improves the classification accuracy, but also provides model’s interpretability. Experimental results show that this approach can gain 64.08% weighted accuracy and 56.41% unweighted accuracy for four-emotion classification in IEMOCAP dataset, which outperform previous results reported for this dataset.

Architecture in an Age of Augmented Reality: Applications and Practices for Mobile Intelligence BIM-based AR in the Entire Lifecycle

Abstract: This paper explores the opportunities for mobile Augmented Reality based on BIM across the planning, design, construction, and operation and maintenance phases of architectural practice. It identifies that high-level uses for mobile Augmented Reality across these phases will extremely helpful to achieve the overall efficiency in AEC industry. Meanwhile, written primarily as an introduction to the development of a mobile intelligence Augmented Reality system for our designers and workers on the construction phase, the paper also analysis important aspects of the system we designed and provide solutions to lightweight BIM data and 3D registration.

Study on Algorithm of Fire Detection Based on Fusion of Multiple Features

Abstract: The paper introduces the algorithm of fire detection based on fusion of multiple features for the early warning of fire. Firstly, on the basis of the color characters of fire image, the color segmentation is achieved under the Ohta color space by the Otsu method. Then, many features in flame area are extracted, which include number of corner points, circularity, area change percentage and height variation. Finally, the features can be fused using the BP network. The results show that the algorithm has characterized by its simplicity and good applicability. Therefore, the fire detection can be done effectively. INTRODUCTION With the development of the techniques of information, the fire detection and pre-warning are moving forward to visualization and intelligence. In a word, the flame can be discriminated by its features, which can be extracted by the technique of digital image process combined by pattern recognition[1]-[3]. These methods not only can realize the early-warning of fire, but also have the advantages including intelligence, anti-interference and high security, which provide new way for fire detection as well as overcoming the shortcoming of the traditional ways. The paper introduces the algorithm of fire detection based on fusion of multiple features in video surveillance system. Firstly, the videos from the surveillance system would be converted into the frame sequence and the color features of flame image can be analyzed. Thus, a new method of color segmentation is proposed by Ohta color ________________________ Hongmei Yan. College of Communication and Information Engineering, Xi’an University of Science and Technology, Xi’an, China

Preliminary Theory of Set DR of Fuzzy Time Series Forecasting Model

Abstract: In order to solve the question which the existing fuzzy time series forecasting model prediction accuracy is not high, this paper proposes the fuzzy time series forecasting model based on differential collection of SD. The general elements expressed in SD( ). Prove: IF a time series forecasting method requires AFER

Privacy-area Aware All-dummy-based Location Privacy Algorithms for Location-based Services

Abstract: Location-Based Services (LBSs) have become one of the most popular activities in our daily life. With the rapid advance of LBSs, there are more threats to users’ privacy. For this reason, while enjoying the convenience provided by LBSs, we have to protect our location privacy. In this paper, we propose two all-dummy-based location privacy algorithms to achieve k-anonymity for privacy-area aware users in LBSs. Different from previous work, on the client side, our method can prevent the center attack and border attack through transforming the actual location to an anchor and the request we send to the LBSs doesn’t contain the actual location of the user. On the server side, we use a rough query before precise query to reduce the processing time and transmission bandwidth. It only returns the results that the client needs. Evaluation results show that our methods can provide more effective privacy protection and lower computation and communication cost.

The Application of Virtual Reality in Education

Abstract: VR, virtual reality technology, is quite popular these days. Virtual realities artificially create sensory experience, which can include sight, touch, hearing, and smell. It has had a profound impact on education, changed some of the past teaching ideas and teaching models. This paper summarizes the advantages of virtual reality technology in teaching and its application in school education, discusses the problems of virtual reality in education and puts forward the corresponding countermeasures. Our research shows that with the guidance of modern teaching theory, virtual reality can make the learners learn actively and happily and improve the teaching efficiency.

Image Retrieval Based on Fused CNN Features

Abstract: Recently, various convolutional neural network (CNN) models have demonstrated their powerful ability as a universal representation for different image recognition tasks. In this paper, image retrieval with different kinds of CNN is considered. Firstly, we extract three kinds of CNN features with the current popular pre-trained CNN models to process image retrieval, respectively. Then, we compute weighted average of the similarity scores of these CNN features and propose an image retrieval algorithm based on the fused CNN features. Extensive experiments on two publicly available datasets well demonstrate that the proposed algorithm is clearly better than the retrieval algorithms based on individual CNN features and other current image retrieval algorithms.

Deep Convolution Neural Networks for Automatic Eyeglasses Removal

Abstract: The facial image under eyeglasses occlusion can degrade face recognition performance. Inspired by the success of deep convolutional neural networks (DCNN) on super resolution, in this paper, a method based on deep convolutional neural network is developed for automatic eyeglasses removal from frontal facial images. To remove eyeglasses on facial images, the proposed approach applied deep convolution neural networks (end-to-end DCNN) to reconstruct the eyeglasses region. We adopt the deep convolutional neural networks (DCNN) approach is designed and trained to learn the mapping between pairs of face images with or without eyeglasses from a large face database in video surveillance. The extensive experiments show that the proposed algorithm can effectively remove eyeglasses, and also can keep the stability of face recognition under eyeglasses on occlusion.

Knowledge representation of entity attribute frame for natural language understanding

Abstract: The key problem in the construction of the semantic knowledge base for natural language understanding lies in the connection between knowledge, language and computation. It’s hardly to get an organized knowledge base without considering the existence modality of knowledge in language and how the knowledge used for the computer. This paper discussed the function and abstract form of semantic knowledge in language comprehension by analyzing the relationship between language and knowledge, and defined the unit and the association of in the way of human understanding of the world. This paper proposed a semantic knowledge representation system which is separated from language and knowledge organization form. This system would enhance many applications in the field of computer natural language understanding.

New Labellings Based on the Idea of Graphical Construction Plus Number Theory for Designing Cryptographs

Abstract: Based on the thought of new password design of “graphical construction plus number theory” as well as the edge magic graceful labelling, we defined a new graph labelling called (ki) 1 m -edge magic graceful totally labelling (m2), and obtain the properties of these two labellings on graphs. We show several new algorithms for constructing graphs having these two labellings, and furthermore show that every graph admits at least a (ki) 1 m -edge magic graceful totally labelling.

The Decision Statistics of the Gaussian Signal against Correlated Gaussian Interferences

Abstract: We introduce an approach to obtaining the approximations of the decision statistics of the fast-fluctuating Gaussian random signals. By means of the presented procedure, there can be found a new much simpler expression for the logarithm of the functional of the likelihood ratio of the Gaussian random process against white noise and correlated Gaussian interference. As a result, new simpler circuits for the receivers of random information process can be synthesized.

A Survey of Network Slicing in 5G

Abstract: The traditional network is facing many problems, making its development in the bottleneck period. Data forwarding efficiency and resource utilization rate are very low, the structure is also complex, and the redundancy is not optimized[1]. In order to solve this problem, communication operators continue to develop communication technology, which is the fifth generation of mobile communication technology, then has been introduced the concept of network slicing, putting the network function virtualization and software defined networks into the network slicing[2]. The new network slicing technology in this paper can improve the network service capability and meet the needs of multi service development.