The paper firstly analyze cache replacement strategies at present, and proposed the ideas of the semantic query cache replacement based on user access features, and describe the semantic similarity calculation and rea...The paper firstly analyze cache replacement strategies at present, and proposed the ideas of the semantic query cache replacement based on user access features, and describe the semantic similarity calculation and realize the algorithm of replacement strategy. The strategy use semantic to match information in the query cache, through dynamic analysis and tracking three characteristics of user access time, user access to content and Business Association, give out the similarity minimum of the cache item, to improve the hit ratio of the cache and the response time and throughput of the server is improved.展开更多
Image fusion can be performed at different levels:signal,pixel,feature and symbol levels.Almost all image fusion algorithms developed to date fall into pixel level.This paper provides an overview of the most widely us...Image fusion can be performed at different levels:signal,pixel,feature and symbol levels.Almost all image fusion algorithms developed to date fall into pixel level.This paper provides an overview of the most widely used pixel-level image fusion algorithms and some comments about their relative strengths and weaknesses.Particular emphasis is placed on multiscale-based methods.Some performance measures practicable for pixel-level image fusion are also discussed.At last,prospects of pixel-level image fusion are made.展开更多
A Mechanism-Inferring method of networks exploited from machine learning theory caneffectively evaluate the predicting performance of a network model.The existing method for inferringnetwork mechanisms based on a cens...A Mechanism-Inferring method of networks exploited from machine learning theory caneffectively evaluate the predicting performance of a network model.The existing method for inferringnetwork mechanisms based on a census of subgraph numbers has some drawbacks,especially the needfor a runtime increasing strongly with network size and network density.In this paper,an improvedmethod has been proposed by introducing a census algorithm of subgraph concentrations.Networkmechanism can be quickly inferred by the new method even though the network has large scale andhigh density.Therefore,the application perspective of mechanism-inferring method has been extendedinto the wider fields of large-scale complex networks.By applying the new method to a case of proteininteraction network,the authors obtain the same inferring result as the existing method,which approvesthe effectiveness of the method.展开更多
文摘The paper firstly analyze cache replacement strategies at present, and proposed the ideas of the semantic query cache replacement based on user access features, and describe the semantic similarity calculation and realize the algorithm of replacement strategy. The strategy use semantic to match information in the query cache, through dynamic analysis and tracking three characteristics of user access time, user access to content and Business Association, give out the similarity minimum of the cache item, to improve the hit ratio of the cache and the response time and throughput of the server is improved.
基金the National Natural Science Foundation of China (Nos. 60775022 and 60705006)
文摘Image fusion can be performed at different levels:signal,pixel,feature and symbol levels.Almost all image fusion algorithms developed to date fall into pixel level.This paper provides an overview of the most widely used pixel-level image fusion algorithms and some comments about their relative strengths and weaknesses.Particular emphasis is placed on multiscale-based methods.Some performance measures practicable for pixel-level image fusion are also discussed.At last,prospects of pixel-level image fusion are made.
基金supported by the National Natural Science Foundation of China under Grant No. 70401019
文摘A Mechanism-Inferring method of networks exploited from machine learning theory caneffectively evaluate the predicting performance of a network model.The existing method for inferringnetwork mechanisms based on a census of subgraph numbers has some drawbacks,especially the needfor a runtime increasing strongly with network size and network density.In this paper,an improvedmethod has been proposed by introducing a census algorithm of subgraph concentrations.Networkmechanism can be quickly inferred by the new method even though the network has large scale andhigh density.Therefore,the application perspective of mechanism-inferring method has been extendedinto the wider fields of large-scale complex networks.By applying the new method to a case of proteininteraction network,the authors obtain the same inferring result as the existing method,which approvesthe effectiveness of the method.
