Abstract With the deepening of application Internet Things in education, its role promoting education and teaching has become increasingly obvious, but corresponding is that remained experimental stage, can not be applied on a large scale. Therefore, according to characteristics educational Things, this paper designs completes new resource database management system Things. The consists three parts: core layer, access layer performance layer. It good expansibility shields complexity...

Attributes in datasets are usually not equally significant. Some attributes unnecessary or redundant. Attribute reduction is an important research issue of rough set theory, which can find minimum subsets with the same classification effect as whole dataset by removing redundant attributes. We use Chi-square statistics to evaluate significance condition It reduce search space attribute and improve speed reduction. Conditional entropy relative adopted a heuristic function. Two decision table...

This paper proposes VDStream, a new effective method, to discover arbitrary shape clusters over variable density data streams. The algorithm can reduce the influence of history and effectively eliminate interference noise data. When streams changes, VDStream dynamically adjust parameters find precise clusters. Experiments demonstrate effectiveness efficiency VDStream.

Enterprise information systems (EIS) play an important role in business process management. Process mining techniques that can mine a large number of event logs generated EIS become very hot topic. There always exist some deviations between model and logs. Therefore, needs to be repaired. For the with selection structures, accuracy existing methods is reduced because additional self-loops invisible transitions. In this paper, method for repairing Logical-Petri-nets-based models structures...

Aspect Sentiment Triplet Extraction (ASTE) aims to extract aspect terms, sentiment polarity and opinion terms explaining the reason for from a sentence in form of triplets. Many existing studies model context by graph neural networks learn relevant information generated graphs. However, some sentences may have syntactic errors or lack significant grammar, which lead poor results on dataset model. In this paper, we propose Fusing Semantic Syntactic Information (FSSI) model, incorporates both...

Collaborative filtering is one of the most successful and widely used recommendation technology in E-commerce systems. However, existing collaborative algorithms face severe challenge sparse user ratings real-time recommendation. To solve problems, a algorithm based on topic proposed. It divides raw rating matrix into many sub-matrixes forms clusters parallelly to reduce data sparsity. Time-based weight acceleration-based are proposed dynamically reflect change interests. The experimental...

To solve the problem of evaluating profile error surface, theoretical surface was built by interpolating design points at method bicubic Non-Uniform Rational B-Spline(NURBS). Measuring were gained laser measurement, and mathematical model for computing error. The particle swarm optimization (PSO) applied to compute minimum distance from measuring which can evaluate accurately. At same time, MATLAB software used realize visualization evaluation free-form surface. Experiments show that...

Abstract Microblog is a popular social network in which hot topics propagate online rapidly. Real-time topic detection can not only understand public opinion well but also bring high commercial value. We design method for real-time microblog data analysis order to detect long lasting events as emerging events. Firstly, mining frequent items algorithm on stream proposed count approximate word frequency. This find the words some time. Secondly, windows size of monitored adjusted dynamically...

This paper proposes a method based on Lossy Counting to mine frequent itemsets. Logarithmic tilted time window is adopted emphasize the importance of recent data. Multilayer count queue framework used avoid counter overflowing and query top- K itemsets quickly using index table.

The microscopic pore structure is the base platform of pore-scale percolation mechanism research. This paper mainly introduce a proposed algorithm for generating stochastic network which can be used to represent he space rock with given input and throat size distributions connectivity ¨C these obtained from an analysis pore-space images. We adopt Mersenne Twister random number generator generate network, spatial correlations are considered in this paper. Quasi-static two-phase model also...

According to the characteristics of high precision and massive amounts data processing during real-time network forensic, combining defects traditional Apriori algorithm which scan sets more times, paper improved algorithm, set is divided into parallel blocks, then use dynamic itemsets counting method weight each block construct tree, depth-first search mark out block, evaluation all items has counted in order acquire frequent itemsets, reducing number scanning, capability forensics,...

The independency between two attribute subsets can be verified based on Chi square statistic to reduce candidate sets. Based this measure, heuristic algorithm employing information entropy for reduction of decision systems is presented by combining rough sets and statistics. And the validity analyzed.

The form error of turbine blade impacts on the work performance directly. To inspect error, cross section curve is constructed as NURBS by interpolating design points in paper, then mathematical model developed searching for four control minimum distance between measure and curve, particle swarm optimization (PSO) applied solving curve. Comparing with BFGS DFP, experiment results show proposed method effective inspecting error.

With the rapid development of Internet, people are confronted with information overload. Many recommendation methods designed to solve this problem. The main contributions proposed in paper as follows: (1) An improved collaborative filtering algorithm based on user clustering is proposed. Clustering performed according similarity user-item rating matrix. So search space reduced. (2) Considering factor that user’s interests may dynamically change over time, a time decay function introduced....