Browsing by Author "Chang, Chan-Chine"

Now showing 1 - 6 of 6

Design and implementation of an intelligent manufacturing execution system for semiconductor manufacturing industry
(2006)
Chen, Ruey-Shun
;
Tsai, Yung-Shun
;
Chang, Chan-Chine
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
This paper proposed an intelligent MES by integrating manufacturing execution system, data warehouse, online analytical processing and data mining system. The data warehouse for this system is provided by the massive amounts of data gathered from a MES. Through some process of integrating MES with data warehouses, data warehouses with decision analysis, and decision analysis with data mining systems. A three-tiered web-based systematic framework has been established. The result of this study is the integration of the MES and data mining system. According to experimental analysis, the product yield and the manufacturing cycle time all have been improved for manufacturing industry.
Ontology-based knowledge extraction - A case study of software development
(2006)
Chen, Ruey-Shun
;
Chang, Chan-Chine
;
Chi, Isabel
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
It is very important that construct a common library from the software development process in a standard analysis pattern. Reusing analysis knowledge in the same domain and the common library can help to build up high-quality applications in limited developing time. System analysis patterns enable a software application to model a specific problem by representing some domain classes and their relationships as modeling components. In this paper, a conceptual model was developed that emphasized the role of analysis pattern and domain library for application with ontology including domain model and dynamic characteristics of classes in tree maps. The result shows two conclusions were derived: first, less development and requirement changes the process costs with common library in ontology and we easy to search and reuse the library components. The second, it is more highly user satisfaction due to the system flexibility and more patients were paid for requirement changes.
Using Bayesian networks to build data mining applications for a semiconductor cleaning process
(2007)
Chen, Ruey-Shun
;
Chang, Chan-Chine
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
Data mining is part of the knowledge discovery process that offers a new way to look at data. It consists of the nontrivial extraction of implicit, previously unknown, and potentially useful information from data. It is the discovery of patterns that lie hidden among a set of data. A Bayesian networks is a specific model of knowledge network. It offers useful information about the mutual dependencies among the features in the application domain. Such information can be used for gaining better understanding about the dynamics of the process under observation. Manufacturing process monitoring using Bayesian networks is a quite useful example. In this paper, we provide Bayesian networks to extract knowledge from data. We present an approach for the Bayesian networks to implement a data mining task for computer integrated manufacturing (CIM). It could find the cause factors in various parameters that affect in semiconductor cleaning process.
Using data mining technology to build an intelligent manufacturing system for semiconductor industry
(2006)
Chen, Ruey-Shun
;
Chang, Chan-Chine
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
In this paper, we propose a flexible three-tier architecture for data mining in semiconductor manufacturing. The architecture allows a high-performance implementation of data mining techniques using association rule. The prototype is able to mine patterns from large data sets that reside in databases. There are benefits to use the data mining control system that is implemented by association rule in computer integrated manufacturing (CIM) environment. The result shows that the proposed intelligent manufacturing system can improve product yield rate and reduce cycle time in semiconductor manufacturing. It could find the best machines of process machine sets that affect processes. It also enhances wafer yield accuracy, reduces production costs and promotes the enterprise competitiveness.
Using data mining technology to improve manufacturing quality - A case study of LCD driver IC packaging industry
(2006)
Chen, Ruey-Shun
;
Yeh, Kun-Chieh
;
Chang, Chan-Chine
;
Chien, H. H.
;
資訊管理與財務金融系註：原資管所+財金所
In recent year, because of the professional teamwork, to improve the qualification percentage of products, to accelerate the acknowledgement of product defects and to find out the solution, the LCD driver IC packaging factories have to establish an analysis mode for quality problems of product for more effective and quicker acquisition of needed information and to improve the customer's satisfaction for information system. The past information system used neural network to improve the yield rate of production. In this research employs the star schema of data warehousing as the base of line analysis, and uses decision tree in data mining to establish a quality analysis system for the defects found in the production processes of package factories in order to provide an interface for problem analysis, enabling quick judgment and control over the cause of problem to shorten the time solving the quality problem. The result of research shows that the use of decision tree algorithm reducing the numbers of defected inner leads and chips has been improved, and using decision tree algorithm is more suitable than using neural network in quality problem classification and analysis of the LCD driver IC packaging industry.
Using data mining technology to solve classification problems - A case study of campus digital library
(2006)
Chang, Chan-Chine
;
Chen, Ruey-Shun
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
Purpose- Traditional library catalogs have become inefficient and inconvenient in assisting library users. Readers may spend a lot of time searching library materials via printed catalogs. Readers need an intelligent and innovative solution to overcome this problem. The paper seeks to examine data mining technology which is a good approach to fulfill readers' requirements. Design/methodology/approach - Data mining is considered to be the non-trivial extraction of implicit, previously unknown, and potentially useful information from data. This paper analyzes readers' borrowing records using the techniques of data analysis, building a data warehouse, and data mining. Findings - The paper finds that after mining data, readers can be classified into different groups according to the publications in which they are interested. Some people on the campus also have a greater preference for multimedia data. Originality/value - The data mining results shows that all readers can be categorized into five clusters, and each cluster has its own characteristics. The frequency with which graduates and associate researchers borrow multimedia data is much higher. This phenomenon shows that these readers have a higher preference for accepting digitized publications. Also, the number of readers borrowing multimedia data has increased over the years. This trend indicates that readers preferences are gradually shifting towards reading digital publications.