Advanced Materials Research Vols. 926-930

Abstract: In order to solve the problem of ignoring items’ attributes in single criteria collaborative filtering recommendation algorithm, a collaborative filtering recommendation algorithm based on multi-criteria decision making was proposed. Multi-criteria decision making was used during the computing of users’ similarity in the algorithm. By averaging all the similarities of sub evaluation with their weight, the algorithm chose the nearest neighbors. Finally, the algorithm used several nearest neighbors to get the recommendation results. Experiments show that the algorithm can find the nearest neighbors more similar after detailing evaluation, and make better predictions for the target users.

Abstract: A compression method for good visual quality is proposed in this paper. It is based on JPEG-XR. The visual quality is measured by structural similarity (SSIM). By some image compression experiments, the minimum SSIM is determined in a manual way. It represents the required visual quality. The image activity measure (IAM) is selected as an image feature. The relationship between IAM and image compression performance under different quantization parameters (QPs) is analyzed. Based on this relationship, proper QPs can be chosen for different images. Then JPEG XR can be used to compress images with these QPs, which will result in good visual quality with fewer bits consumed. The experiments have verified the effectiveness of the proposed method. How to choose a suitable image feature may be our future work.

Abstract: This paper Based on the analysis of the characteristics of image information and image information technology, analyzes some field demand for image information technology, analyzed the image application of news communication management, Analysis of domestic and foreign research present situation, proposed to increase the image signal, research ideas application depth and the breadth of color and its technology in the field , to accelerate the process of social normalization.

Abstract: Clustering analysis is widely used in data mining, e-commerce, graphic processing, bioinformation and text classification. Multicore computing based on CUDA and GPU is one of the new techniques of data processing, which became an active research direction of parallel computing in recent years. After the analysis and serial k-means algorithms, we propose a new compact parallel k-means algorithm which fit for GPU computing and present three main optimization method.The experiment results show that the algorithm is simple, fast and scalable for real-world data processing with comparison of existing other research.

Abstract: This paper mainly make stitching technology of character in scrapped paper of printed file carry on research, which puts forward a stitching method in the light of character and image with multi-feature information in scrapped paper, mainly considering stitching same line character generally possesses similar horizontal projection region and projection center, it adopts human-computer interaction to semi-automatic stitching. First, we make image of scrapped paper binarization and small area filtering (with the exception of the boundary), after characters clustering, we utilize matching degree of pixels information of the left and right edges to make vertical stitching. This method can better tackle stitching and restoration of scrapped paper characters whose edge with segmented character. The scraped papers provided by the national undergraduate mathematical modeling contest were conducted experiment; the results indicate that this method is effective, with strong real-time property, robustness.

Abstract: Co-authoring a manuscript usually connotes a strong influential connection between researchers. Based on data extraction, build the co-author network which maps 511 nodes connections, the nodes represent authors and the links represent the connections with one another coauthors. Within the network, we use software UCINET to analyze its properties. Furthermore, we apply the grey relational analysis method into measuring the author influence and impacts, after extracting five indexes which are the degree, closeness, betweenness, eigenvector and the number of joint publications. By combining network topology and the algorithm ,it will be much better to tap and identify information between authors. Furthermore, it can provide the development of science and technology management and technology with some important guidance.

Abstract: The emergence of blog hot topic means that the user's interest ,participation behavior and various media report coverage reach to its climax,a detecting method of topics on blog based on blog bursty words is proposed. It includes the use of word similarity measure and text clustering analysis which is combined with design strategy in specific period, the use of the main idea of the sudden vocabulary hot topic detection algorithm has to be used and improved in order to generate the final clustering. The experimental results show that the algorithm can obtain an accurate blog topic detection results.

Abstract: The basic principle, training methods and model selection of support vector machine (SVM) are expounded in this paper, and then we introduce SVM in transformer fault diagnosis which can overcome the problems of the structure selection of neural network.

Abstract: Quantitative design focuses on drugs between biological activity and structure parameters of quantitative change rule, so as to apply these rules to guide the design and synthesis of new drugs to predict unknown compounds of biological activity, agent theory and inference mechanism of drugs. This paper briefly introduces the concept of quantitative drug design and computer graphics and its typical applications in pattern recognition, quantitative drug design, and introduces a quantitative drug design system based on pattern recognition, finally will point out their application prospects and some problems to be solved. Quantitative drug design is of great significance for the diagnosis of the disease.

Abstract: This paper has proposed a laser scanning point cloud mosaic scheme on the basis of 2D images matching and 3D correspondence feature points refining. This scheme is aimed to solve existing problems in laser scanning point cloud mosaic technology, such as low efficiency,poor accuracy and low automation. Firstly, the 2D images were generated from the derivative information by an interpolation algorithm. Then 2D correspondence feature points were obtained through GPU acceleration SIFT image matching and eliminates gross errors. In order to improve the accuracy of mosaic,the correspondence feature points must be refined through further constraint. Secondly, the 3D correspondence feature points were acquired based on an inversion algorithm using the 2D correspondence feature points.