Abstract:The present prediction methods for SO₂ concentration suffer from the disadvantages that there is no uniform understanding of pollutant sources and influencing factors, small sample data is sensitive, and prediction methods are easy to fall into local optimum etc. In order to solve these problems, a method for the prediction of SO₂ concentrations on expressway is proposed which is based on fuzzy time series and support vector machine (SVM), and provides a reliable theoretical support for building the highway environmental health monitoring system. Based on the seasonal variation of SO₂ concentrations, the method takes the season as time series, 24h for graining window width. Through the Gaussian kernel function to extract the eigenvalues of the original sample data, which are input support vector machine (SVM) model for training, and k-fold cross validation method combined with the grid division is used to optimize model parameters. Finally, a SO₂ concentrations prediction model is established with the method in this paper. By using 1h average SO₂ concentrations as sample data which are obtained by Shanxi taijiu expressway monitoring station from April 2014 to March 2014, the LIBSVM tool is used to realize the calculation process on the MATLAB platform. The results show that based on fuzzy time series and support vector machine (SVM), the forecasting methods of SO₂ concentration is not restricted by the research of machine rational theory, and supports small-sample learning, otherwise, the nonlinear fitting effect is perfect, and the ability of generalization is well.

Abstract:An immune particle swarm algorithm based on adaptive search strategy is proposed in this paper. Based on the traditional immune particle swarm algorithm, the sub populations are grouped on the fusion algorithm in parallel form, the size of each group is adjusted dynamically, and the search range is also adjusted, according to the maximum concentration of particles. Firstly, combing with the adjustment mechanism of concentration and the maximum value of concentration, the algorithm adjusts the number of sub populations, in order to make full use of the particle source. At the same time, the inferior sub-populations are vaccinated, and the maximum concentration of the particles is used to control the search range of the vaccine. Avoiding the degradation of population, the convergence accuracy and the global search ability of the algorithm are improved. A vehicle scheduling model of open-pit mine is established and simulation experiments are carried out. The simulation results show the proposed algorithm makes full use of the tramcar source, and has certain advantage and good engineering application value.

Abstract:Aiming at characteristics of complexity and nonlinear in enterprise financial data, this paper puts forward a financial early warning model based on General Regression Neural Network which is optimized by Logistic chaos mapping Variable step size Fruit Fly Optimization Algorithm. Firstly, the Logistic chaos mapping is used in Fruit Fly Optimization Algorithm to modify the initial value. Secondly, based on optimal initial value, we modify the step size of FOA in order to find the best Spread. Finally, we analyze the forecast data and select representative indicators. LVFOA shows better ability of global optimization and fast convergence, and it improves the prediction accuracy of GRNN. The simulation results show that the warning accuracy of new model is higher than GRNN model and FOA-GRNN model, better fitting the complex financial data.

Abstract:It is an essential means to detect and analysize the abnormal network traffic in network supervision. And it is also an important research topic in the field of network security management. At the beginning of this paper, we discuss some types of abnormal network traffic, and point out some problems while using traditional anomaly detection methods in network traffic anomaly detection. And then, we specify the latest research achievements of anomaly detection method based on entropy theory which apply information entropy, relative entropy, and active entropy theory to detect abnormal network traffic. Finally, we conclude some problems of anomaly detection methods based on entropy theory and the direction of improvement.

Abstract:In recent years, overload and overdraft phenomenon is more and more serious in inland waterway, and many traffic accidents have occurred from time to time, such as stranding, collision, etc. The corresponding management approach is simple, although the relevant departments have invested huge human, material and financial resources, but the effect is not satisfactory. Based on laser scanning and video analysis, an automatic overload and draft monitoring system is designed, and it is applied in the water areas of Zhenjiang local maritime bureau. The result shows that the system performs excellently in accuracy and real-time etc. in application. And this system can meet the needs of the daily supervision of maritime department.

Abstract:Compared to the host centralized architecture, the distributed architecture based on X86 and cloud computing, has become the mainstream architecture solution, with its characteristics of high extendibility, low coast and high availability. Based on the distributed architecture dedicated to database layer, this article puts forward the solution suitable to the bank core business system, such as database shading, read/write splitting, data sharing, address optimization, high efficient operation and maintenance, etc., which has achieved success in practical application.

Abstract:By using multiple micro controller computers and personal computer, this paper realizes campus express goods management system, which uses the RFID (Radio Frequency Identification) technology to substitute the traditional barcode technology, improving the efficiency of data acquisition. This system has the advantages of fast response, strong real-time and simple operation, and it provides a more efficient solution for package in-out warehouse, inventory, and package position guidance. Experimental results show that RFID technique can improve the efficiency of package storage, to achieve rapid inventory and guide courier parcel position. It can solve the problems caused by frequent changes of goods locations, simplify work flow and reduce labor costs.

Abstract:At present, the distributed business application logs are stored in the local log files on distributed servers, and there is no centralized storage and management, which leads to slow positioning speed for business system problems, and low efficiency in solving problems. This paper provides a distributed log collection and analysis scheme based on OSGi technology. It uses a centralized log storage server to store the log, and provides a set of general log model, so that distributed business application nodes can send the log data to the servers, based on the model. The log storage server receives log data of each node and then unifies storage and interface of analysis display, helping developers to quickly locate and analyze the problem. The scheme is deployed to the application system in the form of the OSGi plug-in, and log is stored in the original way after unloading the OSGi plug-in. Application results show that when the log management scheme is applied to 1000 concurrent business applications which perform logging, the access performance is improved to 2 seconds, and there is no log data loss. According to the developers' feedback, the error logs are clearer, the time of locating problems is obviously shorter than ordinary log storage.

Abstract:An environment intelligent monitoring module is developed and designed under transplanted Linux system, based on the platform with ARM11 core-based S3C6410 as the main controller. This paper mainly develops intelligent monitoring interface based on Qt. With the serial communication between modules and sensors, users can obtain the parameters such as temperature, CO₂, humidity and illuminance etc. in crops greenhouses. By using double buffer technique and pixel mapping, curve display controller is designed to reflect the global and local change trend of all kinds of parameters in the real time. Through developing a character device driver program, this module can control the relay equipment to keep environment indexes to the appropriate standards. Finally, the monitoring web page is designed and published on the server by ASP.NET technology and asynchronous refresh, with socket communication mode to complete data interaction with intelligent module. The users can remote login the system to ensure the suitability of crop growth environment and the safety of the equipment operation.

Abstract:Aiming at solving the problem of how to show a large number of data in graphical form, so as to convey information more clearly, the ECharts visualization technology is adopted to realize Data Monitoring System. The design of the monitoring system is described in detail and the specific implementation is also described. Thereby, it is very important to show data information in graphic form by using ECharts visualization technology.

Abstract:In order to meet the requirements of real-time behavior correctness and certainty of safety critical real-time embedded system, the paper shows a design and implementation of clock synchronization system with hardware timestamp characteristics of the of ETSEC Ethernet device controller. The real-time embedded operating system ReWorks and IEEE 1588 precision clock synchronization protocol is adopted for the implementation. The synchronization accuracy is less than 100 ns, and the accuracy and precision of the clock synchronization system are tested and verified.

Abstract:In view of the internal threat behavior in enterprise information system, especially the abuse of internal user resource, we propose a real-time detection framework based on Agent, which can find malicious insider threat behavior by comparing identify permissions and abnormal operation behavior. The framework is composed of data acquisition module, detection module, audit module and response module. From 4 aspects of identity authentication, access control, operation audit and vulnerability detection, the function of the detection system is described, and the key technology is introduced in detail. The application example proves that the detection framework implements the functions of user's real name login, behavior detection and post audit, fundamentally prevent malicious insiders to obtain illegal data and provide response and intervention capabilities, improving the security of information system. In the end, we summarize the development trend of the internal threat detection technology.

Abstract:Aiming at the problem of poor automation and low level of information integration in the on-line assembly control system of traditional mechanical products, this paper establishes an online quality control system based on machine vision. Firstly, the whole conceptual model of V-OQCS is established, the structure of machine vision is introduced, and then the visual factor library of the online assembly quality control system is established. At last, taking a gearbox assembly system as the research object, of the V-OQCS application is developed to verify the feasibility and effectiveness of the proposed method.

Abstract:With the lower level of automation and high labor costs in the present fabric industry, it is inevitable that the robot will replace human to work. The article presents the robot fabirc scraping system based on the Kinect. First of all, target fabric image is obtained by making use of Kinect. Then, target fabric can be found out by CamShift algorithm and the center of mass is gotten. Finally, the coordinates can be transformed into motion coefficient with robot kinematics to make robot scraping fabric come ture. The experiment results imply the method which can get three-dimensional coordinates of target fabric quickly has more advantages than visual robot.

Abstract:In this paper, a moving object detection system based on FPGA is designed. It uses the modular design mode and pipeline processing method, and makes full use of the FPGA features of high-speed parallel processing and DDR2 SDRAM large capacity characteristics. The system uses the combined methods of the background difference method with single Gauss background modeling and frame difference algorithm to achieve the detection of moving objects. Practical test results show that the target detection system can effectively achieve the detection of moving objects and meet the needs of real-time, and has good performance.

Abstract:Differing from work logic of traditional vehicles, it is a new trend to design a data acquisition and transmission system which can satisfy the characteristic of green vehicles. Using Samsung S3C6410 as main control chip, a general data acquisition and transmission system is designed, and providing the corresponding wireless data acquisition, transmission and status monitoring system. In the data center, programs running on the servers can effectively monitor vehicles, through analysis and storage of collected data. By collecting and analyzing the huge vehicles data, it can be used to realize the functions such as vehicle parameters optimization and fault diagnosis.

Abstract:The paper focuses on the strategy of improving the large amount of concurrent processing during online course selection, based on the existing network condition of Yunnan University of Finance and Economics. With server load balancing by F5_BIG_IP LTM, it can solve the problem that overload server can't response to the concurrent access of a large number of students in the course of real-time course selection. This architecture can greatly solve the problem of network congestion coursed by concurrency access in the process of actual practice. Load balancing technology provides a reliable and stable architecture for increasing Web service application. And it provides a reference for real-timely dealing with the congestion problem of online concurrent access.

Abstract:Popularity of Internet applications has led to the rapid growth of Internet traffic, which brings great challenges for Network operators, and the performance and scalability of traditional flow monitor system cannot meet the needs of operators. Based on the campus network traffic monitoring, this paper attempts to combine many-core technology with Spark, and presents a high-speed traffic monitor system. Many-core processor part is responsible for high-speed traffic collection, processing and log generation; Spark platform handles the distributed storage of large amounts of log data and its high-speed parallel processing. Web Server is used for data visualization. The DNS traffic of the campus network is taken as the object of monitoring, which verifies the feasibility and extensibility of the system.

Abstract:In order to resolve the contradiction between the speed and the precision of transductive support vector machine, a semi-supervised vector machine algorithm based on information feedback is proposed. The algorithm uses the information of the number of last round, the number of reset, the number of unlabeled samples to adjust dynamically the number of labeled samples, and make a tradeoff between region labeling and pairwise tagging. While the progressive evaluating and dynamically adjusting, it can balance the contradiction between the marking speed and accuracy and reduces the transmission and accumulation of errors. The experimental results on AI data sets and UCI data sets show that the proposed algorithm can improve calculation speed on the premise of ensuring the accuracy of label precision.

Abstract:The common data sparsity in recommendation systems makes the nearest neighbor search is not accurate and lets the search results of the nearest neighbor is too small. This will affect the recommended quality and accuracy of the recommendation system, moreover it is difficult to solve in the traditional collaborative filtering recommendation. To overcome the difficulty of data sparsity in recommendation systems, a novel collaborative filtering algorithm is presented which is based on the combination of trust relationship between users and the similarity of scores of the projects. This algorithm constructs the trust relationship among users by using a directed network graph, which can make up the defect that the user's relationship cannot be accurately measured by the user's similarity. The experimental results show that the proposed algorithm can improve the quality and accuracy of the recommendation system.

Abstract:Video frame-rate up-conversion is one of the common video tampering technologies in the time domain, which achieves the frame conversion of video from a low frame-rate to a higher frame-rate by periodically inserting intermediate frames between two frames in the original video. A detection algorithm based on the periodic properties of optical flow for video frame-rate up-conversion tampering is proposed in this paper. Firstly, the video is converted into images. Then the Horn-Schunck optical flow method is used to calculate the optical flow vector of each pixel in each frame, and the optical flow change rate of the adjacent frames is calculate. Finally, the fast Fourier transform is used for the spectral analysis of the optical flow change rate. And whether the inspected video is altered by frame-rate up-conversion would be determined by judging whether the ratio of the max spectrum amplitude and the mean spectrum amplitude is higher than a certain threshold. Experiment results show that the proposed algorithm not only can accurately identify the video tampered by frame-rate up-conversion, but also improve the robust performance of video compression well, and has certain practical application.

Abstract:Most traditional community detection algorithms always consider single factor. Friends who have close relationship may have different concerns and users who have common concerns may not be in a circle of friends. To solve the problems, this thesis presents a hybrid community detection algorithm HCDA, which takes into account the concerns of the community network nodes, but also consider the topological structure of community network. On this basis, it expands iteratively the community by the community gain value between adjacent nodes to find the real interest groups among friend circles. The experimental results illustrate that compared with other methods the proposed algorithm can find the community more accurately.

Abstract:To solve the problem that the number of hidden nodes in regularized extreme learning machine(RELM) affects classification accuracy, sensitive regularized extreme learning machine(SRELM) algorithm is proposed. Firstly, based on the output of hidden layer activation function and its corresponding output layer weighting factor, the formula of computing the sensitivity for hidden node is deduced by residual between actual value and hidden nodes output. Then different hidden nodes are sorted according to sensitivity. And minor hidden nodes are deleted based on classification accuracy of optimization samples. As a result, SRELM classification accuracy is increased effectively. A case study of MNIST handwritten digit database shows that, compared with common SVM and RELM, time consuming of SRELM is almost the same as RELM, and is obviously lower than SVM. Meanwhile SRELM recognition accuracy for handwritten digit is the highest.

Abstract:Different taxi drivers may have different driving preferences when they are cruising to pick up passengers. In this paper, we study the behaviors of taxi drivers' finding passengers with three recommender algorithms, and then provide the taxi driver with the personalized recommendation based on his preferences to the pick-up locations. First, we use the algorithm based on users and collaborative filter of projects to recommend pick-up locations for the taxi drivers. The algorithm is verified by the accuracy rate, proving the feasibility of the two algorithms. Next, taking into account the time factor which would affect the taxies' pickup behavior, we add the time factor into the two algorithms above. Finally, propose the latent factor model (LFM) that breaks the taxi-pickup matrix into two simpler matrices that will help the analysis of the preferences. The results show the three algorithms can effectively form recommendation, and the LFM has a higher accuracy rate.

Abstract:Most of the traditional classifications algorithms have the same classification cost of all categories, which results in a sharp decline in classification performance when the sample data are unbalanced. As to the problem of unbalanced data classification, we combine neural network with denoising auto-encoder and put forward a kind of improved neural network to realize unbalanced data classification algorithm. The algorithm adds a layer called feature damaged layer between input layer and hidden layer. Thus some redundant feature values are lost, and the unbalance degree of data set is reduced. And the results can be obtained after training model obtains optimal parameters and deals with the classification based on feature. It selects three sets of UCI standard unbalanced data sets for experiment. The results show that the accuracy of the algorithm for small data set classification is improved obviously, but when the data set is larger, the classification effect is lower than some classifier. And the overall classification performance of the proposed algorithm is better than other classifiers.

Abstract:Aiming at the problem that the normal segmentation algorithms is difficult to detect the defects from the surface of magnetic ring with complex texture, this paper proposes a novel edge detection algorithm based on wavelet. Based on the weakness to magnetic ring texture using wavelet, Canny edge detection algorithm is adapted to segment the defect edge for the low frequency part after wavelet decomposition and reconstruction. By the find contours algorithm based on threshold segmentation, the magnetic inner-outer contour is extracted and processed. Through morphological processing and logical operation of the defective edge contours, all defects of magnetic ring are abstracted. After sorting the defects and abstracting the main contours of defects, we can judge the contour closure and fill the contour to get the main defect area. Experimental results show that the algorithm has a positive detection rate of 94.7%.

Abstract:MC/DC is a coverage criterion for the verification of avionics software of Level A, which can massively lower down the number of test cases. This paper focuses on the research of the logical expression with coupling conditions, and studies how to obtain the test case set as small as possible. It proposes two solutions which can be used respectively to solve the problems under the conditions of zero-coupling/weak-coupling and strong-coupling, and related examples are showed. The result shows that flexible usage of the two algorithms can solve the problems of rapid generation of MC/DC test case set for general logic expression.

Abstract:With the development of artificial intelligence, natural language retrieval has gradually become a research focus in the field of information retrieval, and the text similarity calculation algorithm directly determines the quality of search. Based on the analysis of the existing research work, some corresponding similarity calculation algorithms are proposed in statistical information level and semantic level information, and a comprehensive calculation algorithm of similarity is proposed finally. The corresponding experiments verify the effectiveness of this comprehensive algorithm in natural language retrieval.

Abstract:With the development of technology, image application technology becomes increasingly important, and many image application technologies have a different requirement and effect for picture images and graphic images. In order to solve the problem of image classification for picture images and graphic images, this paper proposes a new image classification method based on binary classification, which mixes with three kinds of image features such as color, edge and texture. Through the analysis of image feature values that are used K-means clustering method, we achieve the classification for picture images and graphic images. The experimental results show that the method has a good classification effect for picture images and graphic images.

Abstract:Although there are many advantages in traditional K-means algorithm, the clustering criterion function has poor efficiency on classification of the data set with uneven cluster density. On the basis of weighted standard deviation criterion function, this paper proposes a K-means parallel algorithm which is designed and optimized based on MapReduce programming. And it also increases the convergence judgment. Compared with the traditional K-means algorithm, the designed parallel algorithm has a significant improvement in the aspects of accuracy, speedup ratio, scalability and the convergence of clustering results. It also reduces the probability of misclassification caused by the uneven cluster density, and improves the clustering accuracy of the algorithm. What's more, the optimization effect will be more obvious when it deals with lager data size and more nodes.

Abstract:Aiming at uneven resource distribution in cloud computing and bad distribution effect, this paper distributes resources in improved and colony algorithm and particle swarm algorithm. First of all, improve the inertia weight value of particle swarm algorithm, set the fitness function and select particles at the optimal location, then convert the location of selected particles into the value of ant colony algorithm's initial pheromone and improve ant colony algorithm's selection of pheromone through wolves algorithm. Through simulation experiment, compared with ant colony algorithm and particle swarm algorithm, algorithm in this paper has been significantly improved in time to complete tasks and energy consumption.

Abstract:Monkey algorithm is a new optimization algorithm of swarm intelligent. The algorithm can effectively solve the optimization problems of functions such as linear, nonlinear, nonconvex and complex high dimensional function, etc. Currently, it has been widely studied and concerned by many scholars. In order to further improve the solution accuracy of monkey algorithm, this paper puts forward an improved monkey algorithm. Firstly, uniformly distributed Kent chaotic map is adopted as the initial feasible solution of the algorithm. Then, the descending factor is used as the step size in the climb process of the algorithm. Finally, in the simulation experiment, compared with the existing methods, the results show that the solution accuracy of the proposed monkey algorithm is significantly improved, namely, the proposed algorithm is feasible.

Abstract:Aiming at the shortage of the ant colony algorithm in the solving process of cloud computing task scheduling problem, this paper presents a novel task scheduling method of cloud computing based on improved ant colony algorithm, in order to find the best cloud computing task scheduling scheme. Firstly, this paper analyzes current status of research on task scheduling in the cloud computing, and describes the problem in detail. And then ant colony algorithm is used to solve the problem of cloud computing task scheduling, and the defects of standard ant colony algorithm are improved. Finally the performance of the proposed method is tested on the CloudSim platform. The results show that the improved ant colony algorithm not only can find better scheduling scheme for cloud computing tasks, but also speed up the completion of the cloud computing tasks, which has a certain practical application value.

Abstract:The main purpose of the multiple target tracking is jointly estimating the number of targets and their states from a sequence of observation sets, which has the feature of association uncertainty, detection uncertainty, noise and false alarms. In the view of the data association of traditional multiple target tracking algorithm, the large amount of calculation is hard to achieve, while the PHD filter algorithm based on random sets can avoid the problems mentioned above and can estimate the status directly. At present, there is no closed form of solution for the PHD recursion algorithm. This work shows that when both the target dynamics and birth process are linear Gaussian models, the posterior intensity at any time step is a Gaussian mixture. Therefore, the recursive equation can be derived, which can represent the mean of the posterior intensity in terms of Gaussian components, variances and weights. It is demonstrated by simulation that this algorithm can track multiple targets well under non linear, Gaussian assumption.

Abstract:In broadband power line Multi-input multi-output (MIMO) communication, there are multi-user interference and co-channel interference, which needs an application of a beam forming algorithm to eliminate it. Because the channel state information cannot be accurately acquired, it needs a quantitative feedback for channels, so that the feedback error would cause a decline in performance of the system. Aiming at this problem, this paper proposes a beam forming algorithm, which considers the influences of quantization error of the algorithm. In the premise of ensuring the quality of customer service, this algorithm takes maximizing the system energy efficiency as the optimization goal. The beamforming simulation verifies the effectiveness and robustness of the algorithm.

Abstract:During the development of the Android application, the interaction often occurs frequently between the internal application and applications. So the following aspects which include the application interaction pattern, return path, transmission medium and task stack, etc., are studied to improve the system efficiency of the application. Intent components are an interactive medium between applications. Firstly, we make a detailed description for the main attributes, two ways of starting the target component by explicit Intent and implicit Intent. And then we provide a specific implementation code. It is helpful to improve the efficiency of interaction by choosing different Intent startup modes under the interaction between different applications. On the return path of Activity, this paper puts forward two modes and solutions for the return of task stack and TaskStackBuilder object management. As a consequence, the problems of sequential return path or special path are resolved, which has been verified on multiple types of devices.

Abstract:In order to ensure the reliability of the remote dispatching network and improve the utilization efficiency of optical cable network resource, this paper presents a resource allocation approach based on fault link group separation to solve the problem of high resource cost and block probability of shared Backup Path Protection approach. In the construction of protection channel, management of protected resources can be allocated according to the formation of the resource link group. It utilizes the main routing and alternate routing tag to improve the utilization efficiency of backup channel, realizing the saving of the main working route resource. Finally, the proposed method is verified under different network simulation scenario, and the performance is also analyzed in theory.

Abstract:In recent years, with the vigorous development of computer science and wireless sensor network, how to address big trajectory data is becoming a concerned issue increasingly. Because of massive trajectory data, there is an increasing focus on storage and search of big trajectory data. In view of this, based on the document type non relational database MongoDB, we propose a spatio-temporal index of road network which is based on quad-tree. For the 1915 taxis in Taiyuan, the 500,000 pieces of trajectory data are searched. With different data and different number of concurrency, we compare the efficiency of spatio-temporal index with that of MongoDB composite spatio-temporal index. Experimental results show that our method performs well when data volume is larger than 100000. It can adapt to spatio-temporal queries with different number of concurrency, proving that the method is feasible and efficient.

Abstract:Based on the existing data mining technology,this article adopts the method of optimizing the initial clustering center to improve the k-means clustering algorithm, we can study of Xinjiang agricultural University student id card consumption data for the research and analysis,and provide decision support for the related departments.First of all, according to the demand analysis, we will choose some students for school year 2014-2015 real data in one cartoon system as data analysis,and data preprocessing, at the same time, we will choose the dining room number and amount, the supermarket consumption number and amount, the dining place for experimental characteristic attributes;Secondly, we use the improved clustering algorithm to analyze the data, and comparative analysis based on three kinds of distance measure under the k-means clustering algorithm;Then, the analysis conclusion, the student canteen consumption behavior and supermarket consumption behavior;Finally,the study was based on the conclusions of analysis provides decision support for schools.

Abstract:In order to calculate and estimate travel time with the data of video vehicle detectors, data of queue length is applied to the calculation of travel time and the roads are researched with the improved BP neural network algorithm and time series analysis. The decision coefficient is 93.36% when queue length is added to the calculation, which is improved by 41.03% compared with the neural network algorithm for the traffic data only, and 23.37% compared with the BPR algorithm. Using real-time travel time can been used to predict the follow-up travel time. And through the time series analysis, the relative error is 0.06. The average relative errors are 0.14 and 0.15 respectively for forecasting the travel time of the next period and next cycle. Results show that the queue length has higher accuracy for calculating travel time, which can be used to predict travel time of the urban road. The algorithm can provide ideas for calculation of index for other algorithms in the field of intelligent transportation and can also provide decision support for improving the traffic situation.

Abstract:With the rapid development of software technology and network, internet technology has been increasingly applied to the field of education, and internet education has been more and more accepted and adopted by more and more people for its increasingly rich form and content. Online learning has become a new and effective way of learning. With its unique characteristics of the time independent, timeliness, repeatability, online learning can makes every internet user learn online. At the same time, it also presents a challenge to the online learning system that how to make the learning system run stably under high concurrent access. This paper mainly introduces the design and implementation of the learning system based on Dubbo distributed architecture. The system can attract the attention of children through the educational games, improve their learning enthusiasm, and then guide children to learn effectively by network. It can ease the pressure of high concurrent access to the server, through the Dubbo distributed cluster architecture.

Abstract:There are some problems in the identification of poor students in colleges and universities, such as high cost, the lack of credibility and the inconsistency of standards. Through anglicizing the consumption behavior of campus one card solution, we can depict the consumption characterization of poor students. And then we build Markov model for consumption behavior of poor students, and put forward the concept and calculation method of line index. Based on calculating the similarity index for both the consumer behavior of student model and the consumption behavior of poor student model, we can identify the poor students. This method has the characteristics of high computational efficiency, fast speed, low cost and easy data acquisition, which has consistent evaluation criteria in the same school. And the average recognition rate for poor students is than 90%. It can be used as a powerful auxiliary tool for the identification of poor students in colleges and universities.

Abstract:Electroencephalography (EEG) classification is the key point of brain-computer interface application. How to find effective feature is the major issues in EEG classification. Although several effective methods like support vector machines or neural networks have already been applied to EEG classification, but these methods need a large amount of prior knowledge to find the features of the data. Since the brain electrical signal appears to be more susceptible to noise interference and there are wide individual differences, so that effective features are difficult to been found. Meanwhile, it is difficult to improve the accuracy of the EEG classification, especially in the advanced cognitive process in the cigarette craving. In order to solve this problem, we use convolution neural networks (CNN) to classify EEG of cigarette craving patients under different status of cigarette craving. Compared with the traditional method, CNN does not need to manually extract features. It can directly train the original EEG data. More importantly, it can satisfy the demand which is to obtain the real-time feedback in the cigarette craving treatment process for classification results.

Abstract:Compared with traditional single block architecture, microservice architecture has many advantages, such as flexible technology selection, independent deployment, and independent scalability more suitability for the current needs of the internet age, etc. But microservice architecture also introduces new problems such as service registration, service discovery, service fault tolerance. On the basis of the analysis for problems mentioned above, this paper proposes one implementation of microservice framework, which can solve service registration, service discovery, service fault tolerance and other common problems. Based on this, developers only need to focus on the development of business functions, so that it can simplify the difficulty of system development and improve development effectiveness.

Abstract:We design and implement a polymerization system for web learning resources based on CURL. Using the CURL multi-threading functions, the system can send data to all resource websites at the same time through GET or POST method. And then the system can unify all returned HTML codes from resources websites, and use regular expressions to extract the search results area of the returned codes. It uses PHP DOM manipulation class to fix the image and the URL addresses of links in the code, and then loads all returned code into the same page. Thus it can realize piecewise loading by using waterfall flow model.

Abstract:In order to facilitate the scientific research workers to know the impact factor of SCI (Science Citation Index) journal in their own field, the program realizes a filter of more than 11,000 kinds of SCI journals in a certain field that are published annually to get impact factors of the journal that are related to their own subject. SCI journal involves nearly 100 research fields, but each researcher only cares about one or two research areas. At present, the scientific research workers mostly use Excel tool to handle the journal data by hand. Thus it needs a huge workload and the accuracy cannot be guaranteed. The batch data of workbook in SCI journal dataset can be fast processed by VB6.0 (Visual Basic6.0) program, and the procession has the characteristics of high accuracy and high-speed has a wide range of application.

Abstract:A kind of scheme for vehicle data acquisition and uploading is proposed in this paper. And the real-time monitoring is implemented on the basis of data exchange between Bluetooth module of the android platform and OBD module on the vehicle. Moreover, the collected data is uploaded to the web server which is built on JEE. And then the data are showed in the form of charts on the web-side after data are analyzed and processed on the server. The scheme has a lot of advantages such as low cost, strong operability, high efficiency easy maintenance and so on.

Abstract:Aiming at the increased requirement of the healthy management, this paper presents a management App which is based on Android smart phone. This software is based on the platform of Android Studio, which includes such modules as:register module, physical parameters (body weight, heart rate, blood pressure, etc.), input and query module and physical examination knowledge, etc. The system is mainly consists of SQLite database, wireless network, Bmob back-end service platform, and the main technology includes that user register information is stored on the Bmob platform, personal physical exam data can been saved and queried in SQLite database, and physical exam knowledge can been saved in the form of files. Experiments show that this system realizes the management of personal health data, help user record their health status, and study the basic knowledge of physical examination.