Advanced Materials Research Vols. 753-755

Abstract: Automated assessment technology for subjective tests is one of the key techniques of exam systems. A model based on domain ontology is proposed in this paper, which can be used in exam systems to estimate subjective tests. After analysing the present research status of subjective automated assessment technology, the paper makes a study on the construction method of domain ontology by taking software engeering domain as an example. Semantic similarity calculation based on domain ontology is used for automatic assessment in this paper. The automatic assessment system can divide a sentence into a series of phrases by using the natural language processing technology and get the score by evaluating the semantic similarity of the student's answer. The experiments show that the results of the system which has certain valuable feasibility and applicability are credible and the scoring errors are acceptable.

Abstract: Most of the traditional network management system used c/s mode, and deployment is not convenient. When device warns, operators will face a lot of alarms, and unable to determine the cause of the accident quickly. In this paper, intelligent alarm function design based on RIA / Flex for the power communication network has visual user level and deployment convenience. The alarm of network management is classed based on the business features. Mechanisms are presented in the paper for the filtering, detection and display of alarm information and monitoring of the electric power communication devices. We establish intelligent alarm system for power communication network based on RIA technology.

Abstract: This paper presents a novel saliency detection method for color image based on watershed method and regional spatial attention model. Watershed method is improved to divide the original image into several sub-regions. And regional spatial attention model is specifically designed in the terms of visual attention mechanism from human's psychology and physiology. Experiment results show that our method can accurately and quickly detect the salient regions from color images, and compared with traditional ones, our method performs better.

Abstract: According to Wiener-Khintchine theorem, autocorrelation function of an image was proved that can be used as sharpness evaluation function. An image sharpness index k is proposed and the relationship between content of high frequency components of an image and the width d of its autocorrelation function main lobe was constructed through Fourier transformation. The direct ration relationship between the sharpness index k of an image and the width d of its autocorrelation function main lobe was also validated by experiments. Theoretical analysis and experiment results indicate that d increases when an image is blurry and d decreases when the image is clear.

Abstract: Sentiment analysis on micro blog data has attracted much attention recently. In this paper, according to the characteristics of Chinese micro blog, a method based on semantic rules was provided to analysis emotional of micro blog hot topics. This method considers the adverbs of degree features, adverbs of negative features and emoticons features on the basis of emotional words in the micro blog text. Our proposed method is compared with the method based on emotional dictionary on experiment. The experimental results show that the proposed method is feasible and effective.

Abstract: The paper is focused on the capacitated vehicle routing problem. Tabu search algorithm is an algorithm based on neighborhood search. According to the features of the problem, the essay centered the construct initial solution to build neighborhood structure. For the operation, 1-move and 2-opt were applied, it can also fasten the speed of convergence, and boost the search efficiency. Finally, the good performance of this algorithm can be proved by experiment calculation and concrete examples.

Abstract: Double-direction two-dimensional Maximum Scatter Difference (2D2MSD) based on Maximum Scatter Difference (MSD) was proposed,which overcame the small sample size problem of LDA, and data were more concise. In the Weizmann human action database, experimental results showed the algorithm was fast, the average recognition rate reached 92% and the highest recognition rate reached 100%.

Abstract: With the development of domestic petrochemical enterprises, the need of the import equipment is increasing, so the training for the worker is extremely urgent. This paper designs a set of simulation training hardware system, using the current material resources, according to the configuration file free configuration, so as to reduce the development cost, to speed up the development speed and efficiency. At present the simulation system has been used in domestic petrochemical enterprises, and achieved good economic and social benefits.

Abstract: Based on study of users behaviors, interests, hobbies, professional fields and habits, personalized active information service in digital library is a kind of information service which takes the user's personalized needs as the center. Its key technologies includes user modeling technology, intelligent agent technology, ontology technology, Web data mining, personalized recommendation technology and information push technology etc. Based on the above technologies, this paper constructs a kind of multi-technology personalized active information service model, which can give full play to superiority of various technologies, and has the feature of intelligence. The model applies the technology of user modeling, ontology, intelligent agent, and draws on the general principle of personalized active information service, which is characterized by self-learning, self-renewal and other intelligent elements.

Abstract: In this paper, we propose an algorithm for cluster generating and service discovery based on service popularity in wireless community network,which makes full use of the imbalance of the service request in the network, leading to that the requests could be replied quickly in shorter time. At the same time, the algorithm could adapt well to the wireless environment and incur less update information when the topology is changed. According to the simulation in the OMNET++, our algorithm performs better than the flooding and CRP in several aspects.