Papers by Keyword: Classification

Abstract: In this paper, a new filter network is presented that is based on Radial Base Function Networks (RBFNs). The output layer of the network is modified, in order to make it more effective in certain fuzzy control systems. The training of the network is solved by a clustering step, for which two different clustering methods are proposed. The suggested structure can efficiently be used for data classification.

Abstract: This paper introduces a smartphone-based classification system for an indoor environment of a walking person. The system relies only on smartphone inertial data and it can be considered as a smartphone-based aiding system for an indoor navigation. In addition, it does not need pre-installing of wireless network in the environment or heavily tuning process before the navigation run. Therefore, this system can be used as an aiding block where the person wants to localize himself in an indoor environment starting from known navigations point. This system categorizes person navigation in indoor environment into three types of classes: walking straight, turning right, and turning left. There is an ELM (Extreme Learning Machine)-Based neural network for deciding the class of the current navigation action. The evaluation measure shows that the best performance is obtained with the Radial Basis Function (RBF) as the activation function of the neural network. Also, the obtained accuracy rates up to 95%.

Abstract: For the reasonable grouping of the 16 vessel traffic flow observation lines along Yangtze River to compare them with each other, essential data of nearly 6 years taken with the mean relative growth rate of traffic flow over the yearly same period set as clustering index, the clustering of the 16 vessel traffic flow observation lines was realized on MATLAB in accordance with the set index based on the basic principle and processes of Agglomerative Hierarchical Clustering Method, which indicated that the observation lines can be divided into 3 classes, and particularly Jiujiang took an exclusive class obviously different from the other two. This work would provide scientific basis of grouping the observation lines in order to reflect shipping economy of the regions where the observation lines are located and to provide reference for management decisions.

Abstract: Power quality disturbance detection and identification is the prerequisite and basis for the power quality management and control. This paper presents a new power quality disturbance detection and classification method. Firstly, the time-time transform is applied to power quality disturbance signal analysis. According to spectrum analysis results of the diagonal elements of time-time transform matrix, a preliminary judge about whether the disturbance signal contains harmonics and inter harmonic was given. For disturbances with non-harmonics, based on time-time transform modulus matrix diagonal sequence, the beginning and ending time of the disturbance is located, and the disturbance amplitude is calculated. For the disturbances which contain harmonics, time-time transform is perform twice to get the row mean value curve and the column mean value curve, which are required by disturbance time location and amplitude measurement. Finally, disturbance classification had realized by using rule tree. Simulation results reveal that this method is very robust and adaptable, which can identify transient power quality disturbance with minor magnitude under noisy environment, and the recognition rate is satisfactory.

Abstract: This paper analysis the characteristics of High Spatial Resolution Remote Sensing Image (HSRRSI), consider the drawback of pixel-wise Classification, an Object-based method was taken to classify the image. The paper firstly preprocess and coarsely segment the image, secondly optimize the segmented results for deeply processing, then extracting the image feature from the original image and segmented results, finally classify the image and assess the results.

Abstract: The purpose of this paper was to study the corrosion resistance of Co-Cr and Ni-Cr dental alloys in simulated artificial saliva by means of electrochemical techniques. Co and Ni based alloys are widely used in dental skeletal structures and orthopedic implants such as screws, pins and plater. And recently they have been applied for making stents. The advantages of these alloys include low cost of casting, matching thermal expansion coefficient with the ceramics of metal-ceramic restorations, and acceptable mechanical and tribological properties in vivo. The microstructure of investigated alloys were studied by using optical microscopic observation, X-ray diffraction measurements and Scanning Electron Microscope with X-ray microanalyzer. The mechanical properties were characterized by microhardness tests.

Abstract: With advances in data collection and generation technologies, environments that produce data streams is more and more. In recent years, the network application is further universal and the applications of a single data stream transfer toward a multi-node distributed data streams, such as sensor network, network monitoring, web log analysis and the credit card transaction data of multiple sites. These data is not only real-time, continuous and large scale, but also distributed. How to manage and analyze large dynamic datasets is an important subject that researchers are faced with. In view of the situation, it presented the formalization description of homogeneous and heterogeneous distributed data stream in this paper, analyzed advantages and disadvantages of the centralized stream processing architecture and distributed streaming processing architecture, discussed the recent progress in distributed data stream classification algorithm, summed up the problems and challenges faced by the distributed data stream mining, and possible future research directions.

Abstract: CIn order to build gray level co-occurrence matrix suitable to natural texture, a method based on separable criterion was proposed. Combined correlation matrix of feature parameters with character of natural texture, 5 independent feature parameters are extracted from 11 feature parameters of gray level co-occurrence matrix (GLCM). Building factors of GLCM, which is appropriate to describe wood texture, are confirmed by using the separable criterion, when d equals to 2 and g equals to 16.

Abstract: Through the sheet metal digital information classification, induces the 2D graphics, 3D model, data forms, the information such as process design chart, file information is automatic extracted and meta data description information is interactively supplemented , data is stored in the database records, two kinds of query is designed for recorded data query and keywords retrieval methods.Engineering database system designed to define meta data, storage, retrieval, information coding and classification of four levels, combined with the engineering application of system requirements and function module design,interface of sheet metal unfolding system is developed.The system improve the design capability of sheet metal enterprises, reduce manufacturing cost, enhance the capacity of information management .

Abstract: Nowadays, multi-relational classification has become a hotspot for research and application in the field of data mining. Compared to the single table with simple structure, multi-relational tables is more complicated. However, not all of the information in the tables has good effects on classification. It may decrease the classification accuracy of the algorithm when irrelevant relations are added. In this article, we optimized the multi-relational tables using the usefulness of the backgrounds to remove those relations which have little effect on the classification. The results show that, this method is effective.