Aiming at the problems of the single result of recommending similar users and insufficient understanding of user interests and behavior information for single social network platforms, a User Recommendation method of Cross-Platform based on Knowledge graph and Restart random walk (URCP-KR) was proposed. First, in the similar subgraphs segmented and matched by the target platform graph and the auxiliary platform graph, an improved multi-layer Recurrent Neural Network (RNN) was used to predict the candidate user entities. And the similar users were selected by comprehensive use of the similarity of topological structure features and user portrait similarity. Then, the relationship information of similar users in the auxiliary platform graph was used to complete the target platform graph. Finally, the probabilities of the users in the target platform graph walking to each user in the community were calculated, so that the interest similarity between users was obtained to realize the user recommendation. Experimental results show that the proposed method has higher recommendation precision and diversity than Collaborative Filtering (CF) algorithm, User Recommendation algorithm based on Cross-Platform online social network (URCP) and User Recommendation algorithm based on Multi-developer Community (UR-MC) with the recommendation precision up to 95.31% and the recommendation coverage up to 88.42%.

The main communication mode of graph computing applications is spatiotemporally random point-to-point fine-grained communication. However, existing high-performance computer network systems perform poorly when dealing with a large number of fine-grained communications, which affect the overall performance. The communication optimization in application layer can improve the performance of graph computing application effectively, but this brings great burden to application developers. Therefore, a structure-dynamic message aggregation technique was proposed and implemented, which produced a lot of intermediate points in the communication path by building virtual topologies, so as to greatly improve the effect of message aggregation. By contrast, the traditional message aggregation strategy generally performed only at the communication source or destination with limited aggregation chances. In addition, this proposed technique adapted different kinds of hardware conditions and application features by flexibly adjusting the structure and configuration of the virtual topology. At the same time, the runtime system with message aggregation for graph computing was proposed and implemented, which allowed the runtime system to dynamically select parameters when executing iterations, so as to reduce the burden of developers. Experimental results on a system with 256 nodes show that typical graph computing application performance can achieve more than 100% improvement after optimized by the proposed message aggregation technique.

The existing query analysis methods usually treat the entity resolution as an offline preprocessing process to clean the whole data set. However, with the continuous increasing of data size, such offline cleaning mode with high computing complexity has been difficult to meet the needs of real-time analysis in most applications. In order to solve the problem of aggregation query on duplicate charged records, a new method integrating entity resolution with approximate aggregation query processing was proposed. Firstly, a block-based sampling strategy was adopted to collect samples. Then, an entity recognition method was used to identify the duplicate entities on the sampled samples. Finally, the unbiased estimation of aggregated results was reconstructed according to the results of entity recognition. The proposed method avoids the time cost of identifying all entities, and returns the query results that satisfy user needs by identifying only a small number of sample data. The experimental results on both real dataset and synthetic dataset demonstrate the efficiency and reliability of the proposed method.

In order to reconstruct 3D objects with relief-texture quickly, a modeling method based on real-time interaction for 3D objects was proposed. Firstly, the input object or image that needed to generate relief texture was converted into an initial depth map, then the depth map was converted into a gradient image. The obtained gradient image was compressed and filtered, then the continuous relief depth map was reconstructed by solving the linear equation. Secondly, using the proposed relief-texture-mapping algorithm based on mesh intersection, the optimized relief depth image was pasted on the goal object surface by real-time interaction including translation, rotation and scale under 3D scene directly. Finally, an object with relief-texture was achieved through re-triangulation of the goal model. The experimental results show that the proposed method can quickly generate concave or convex, especially multiple reliefs on the goal object, the final obtained model can be directly applied to 3D printing without any other processing.

Focusing on the issue that extractive summarization needs to automatically produce a short summary of a document by concatenating several sentences taken exactly from the original material. A single document automatic summarization algorithm based on word-sentence co-ranking was proposed, named WSRank for short, which integrated the word-sentence relationship into the graph-based sentences ranking model. The framework of co-ranking in WSRank was given, and then was converted to a quite concise form in the view of matrix operations, and its convergence was theoretically proved. Moreover, a redundancy elimination technique was presented as a supplement to WSRank, so that the quality of automatic summarization could be further enhanced. The experimental results on real datasets show that WSRank improves the performance of summarization by 13% to 30% in multiple Rouge metrics, which demonstrates the effectiveness of the proposed method.

Since the relation of labels in Binary Relevance (BR) is ignored, it is easy to cause the multi-label classifier to output not exist or less emergent labels in training data. The Multi-Label classification algorithm based on Bayesian Model (MLBM) and Markov Multi-Label classification algorithm based on Bayesian Model (MMLBM) were proposed. Firstly, to analyze the shortcomings of BR algorithm, the simulation model was established; considering the value of label should be decided by the attribute confidence and label confidence, MLBM was proposed. Particularly, the attribute confidence was calculated by traditional classification and the label confidence was obtained directly from the training data. Secondly, when MLBM calculated label confidence, it had to consider all the classified labels, thus some of no-relation or weak-relation labels would affect performance of the classifier. To overcome the weakness of MLBM, MMLBM was proposed, which used Markov model to simplify the calculation of label confidence. The theoretical analyses and simulation experiment results demonstrate that, in comparison with BR algorithm, the average classification accuracy of MMLBM increased by 4.8% on emotions dataset, 9.8% on yeast dataset and 7.3% on flags dataset. The experimental results show that MMLBM can effectively improve the classification accuracy when the label cardinality is larger in the training data.

Real networks are of large delay and dynamics. According to the IMC (Internal Model Control) and improved control theory model with TCP/AQM, an active queue management algorithm suitable for the large delay network environment was designed to cope with the large delays. For the dynamics of the networks, the impact that the change of network parameter brings to the algorithm was analyzed to correct the algorithm parameter online. The adaptive active queue management algorithm suitable for the large delay network was acquired. The reliability of the algorithm has been verified by NS2 simulation.

HW/SW partitioning is a key issue in HW/SW co-design of embedded system. In this paper, NSGA-II was applied to HW/SW partitioning. Each run of the algorithm can produce many Pareto-optional solutions. The satisfactory solution can provide an effective tool for measuring the performance of different objective functions, and impove the designing efficiency. The results show that the method can get global optional solutions to fulfill system performance constraint for embedded system.