Applied Mechanics and Materials Vols. 321-324

Abstract: Performance test and fault prediction is the core challenge in building robust cloud computing platform. This paper converted fault prediction problem into a machine learning problem. Based on extracted software feature, software faults were predicted using support vector regression machine. Experimental results show that new method can improve the precision of fault prediction.

Abstract: This paper describes the design and implementation of the Pusat Pengajian Diploma Industrial Training Online System (PiTOS), which was designed to be a user friendly, generic, web-based system to ease accessibility to all vital information related to students’ industrial training in a more organized and efficient way in a paperless environment. The solution we developed is a web application, written using PHP and utilizes the MySQL as the database management system.

Abstract: Considering the difficulty of information collection and integration due to the rapid growth of information, we need an efficient tool to do these jobs. A proposal is be put forward to build a data integration system to collect the source data and preprocess the heterogeneous data and then convert/extract data to the data warehouse. Through experiment and analysis, this paper designed an information process flow and implemented the data integration system, based on B/S framework with the database technology, to deal with the college related information.

Abstract: The main objective of this study is to analyze the relevant factors of nuclear reactor 1 in Fukushima nuclear accident. The analysis results have shown that GM(0,N) is effective and applicable. The forecasting result of the level of radiation 60km away from Fukushima Dai-ichi Nuclear Power demonstrates that GM(0,N) provides very remarkable predication performance compared with traditional multiple regressive.

Abstract: Considering the functional requirements of essential service, value-added service, prediction service and personalized service, which are demanded by users from university, enterprise and government, this paper designed an infrastructure of university information service platform using data warehouse technology. By means of the infomation resource integration method put forward by this paper, the platform realized the subject-oriented, multi-scale service to meet users service requirements and support decisions.

Abstract: Because MOLAP uses Multi-dimensional database (MDDB) to store data, there are limits to implement OLAP, such as data growth. Also, there is a close relationship between the influencing factor of MOLAP data growth and OLAP model, and the models will influence the size of cubes, also the performance and expansibility of OLAP system. This paper studies the approaches of MOLAP data storage organization, aggregate levels of dimension, compound growth factor (CGF), the relationship of query performance and percentage pre-calculated, and giving a brief analysis to how to use CGF to ravel out the data growth of MOLAP.

Abstract: The aim of this paper is to explore dynamic multi-attribute decision making (DMADM) problems in which the decision making information of alternatives is collected at different stages. Firstly, the area closeness degree is applied in normalizing the raw data. Secondly, the weights of different stages are determined by according to the principle of new information priority. The technique for preference by similarity to ideal solution (TOPSIS) is improved to aggregate the information from different stages. Finally, the example is illustrated to demonstrate the practicality and effectiveness of the proposed methods.

Abstract: Real-time data exchange platform is of essential significance to achieving digital mining. This paper, based on a systematic introduction of the relevant technologies and functions of the real-time data exchange platform, proposes its architectural design and concludes with relevant research on the key technology of the system.

Abstract: Physical functions digital archive, principally consisting of basic personal information and evaluation information, is the management process of body functional standard and scientific record, which makes it easy for sports teams and sports research institute to compile and use the digital files. Functional digital archive has a high density of data accumulation and store, transmission of shortcut, information intelligent retrieval, dissemination of intelligence, sharing of information resources and so on. Its requirements in standards, test method, recording method, recording and upload format for uniform and management modes take into account authorization and security.

Abstract: Data mining is to discover knowledge which is unknown and hidden in huge database and would be helpful for people understand the data and make decision better. Some knowledge discovered from data mining is considered to be sensitive that the holder of the database will not share because it might cause serious privacy or security problems. Privacy preserving data mining is to hide sensitive knowledge and it is becoming more and more important and attractive. Association rule is one class of the most important knowledge to be mined, so as sensitive association rule hiding. The side-effects of the existing data mining technology are investigated and the representative strategies of association rule hiding are discussed.