Transcript Document
Extending Mining Applications towards Web Technology in Forest Industry Research&Development Agenda Elena Irina Neaga Forac Research Consortium Laval University, Québec City Canada E-mail: [email protected] Outline • • • • • • • • • • Factors that affect the adoption of DW&DM Customer Relationship Management Supply Chain Management Demand Chain Management Workflow Concept Web and Text Mining Semantic Web Standardization and Integration Issues Research Contributions Future research directions, applications and challenges Factors that affect the adoption of DW&DM The investments on DW&DM technologies are very expensive and the failure is pretty high compared with other IT systems. There is little research regarding the managerial, tactical and strategic aspects of the adoption of DW&DM. The adoption may result based on the real needs of companies and not from the perspectives and conclusions of the research. Pro-active and deep analysis of business problems which require the application of DM. The statistical and optimization methods may be enough. The potential advantages should need to be predicted before obtaining the results. CRM CRM is the process by which forest companies manages their interactions with customers in the same way as other industrial enterprises. Dedicated CRM systems for forest industry, associated tools and methodologies are integrated applications that implement an interface between a specific company and its customers. E-CRM systems overlap B2C interfaces in ECommerce sites. CRM (continued) CRM is also the core activity of e-business, and in the framework of forest products enterprises it could be integrated with SCM and ERP systems as well as other e-marketing applications which may use market basket analysis. Several companies complement their CRM and ERP applications with other Business and Market Intelligence systems. CRM (continued) A CRM system applying DM might be composed of the following sub-systems: •Customer Profiling which is the system that implements the process of discovery of patterns within customer databases which provides new information and knowledge. This system is mainly divided into customer acquisition and customer retention which may also be defined as customer loyalty. •Customer Profitability uses DM in order to understand, optimize and improve it. Customer profitability is also logically linked to customer loyalty. •Customer Segmentation applies DM in order to discover discrete segments in a customer database. •Predicting Customer Behaviour includes churning which represents the process of customer moving from one company to another. Supply Chain Engineering Viewpoint Operational Inventory and Control Production Planning and Scheduling Business Viewpoint Design Integrated Operational Systems Information Sharing, Coordination and Monitoring Strategic Relationship Development Competitive Advantage Demand Chain Management DCM may be defined as the extension of the operations from a single business unit or a company to the whole chain. DCM is a set of practices aimed at managing and coordinating the whole demand chain, starting from the end customer and working backward to raw material and suppliers. The main objectives: the development of a synergy along the whole demand chain. the definition of a focus on specific customer segments and meeting their needs. DM may provide an alternative or a refined solution to the forecasting demand using Bayesian time series [Spedding, Chan, 2000],[Cheung et al., 2001]. Workflow Concept It defines a comprehensive approach for coordinated execution of multiple tasks or activities. Business and production processes modeling and management. Generally WfMSs are for business processes as DBMS for data. Support e-business applications and enterprise integration, collaboration and coordination. Existing Workflow standards defined by WfMC and W3C. Workflow Mining is the discovery processing applied to Workflow systems. Standardisation Issues Existing standards: Related Standards: Predictive Model Markup Language Semantic Web (PMML) Standards (RDF, XML and XMI RDFS, OWL, etc.) SQL/MM Part 6:DM Web services (SOAP/XML, WSDL, Java Data Mining UDDI, etc.) (JDM) Grid services (WSRF, OMG Common OGSI, etc.) Warehouse Metadata (CWM) for DM Research Contribution Framework for Distributed Knowledge Discovery Systems Embedded in Extended Enterprise, Loughborough University PhD thesis: @ 2003, Loughborough, United Kingdom. Standard integration of KD & DM systems in an extended manufacturing enterprise. A unified object-oriented framework for the development of distributed KD/DM systems. KD/DM Products Systems UNIFIED FRAMEWORK Systems for CRM SCM ERP Modeling a Generic DM Application Class Diagram Data Pre_processing_1 Classification Statistical Analysis global Association global Sequential Patterns ApplicationSpecification CRM SCM ERP Market Analysis 0..* Product Life Cycle Production_Inventory 0..n Data Pre_processing_2 Cleaning Profiling Integration Selection Transforamtion opname() <<Data_Mining()>> Data Mining AssociationRule Classification Clustering Statistics Systems_Implementing _Algorithms FuzzyLogic opname() 0..1 Subject-oriented Other Algorithms Visualization 0 Cleaning Profiling Integration Selection OLAP 1 Transformation Dedicated_Systems_for _financial_market PolyAnalyst MyCorbaInterface Applying OMG’s CWM-DM Main Diagram Applying OMG’s CWM-DM Settings Diagram Related Contributions Methodological and standard applications of data, web and text mining systems. Using OMG methodologies, architectures, models and midleware projects such as: UML, CORBA, MDA and CWM. Adhering to the existing reference architectures for enterprise integration and modeling, and ISO standards such as: CIM-OSA, ARIS, PERA, GERAM and RM-ODP. Related Contributions (continued) KD/DM Core Package KD/DM Interfaces CWM/Java C/C++ Corba/IDL C++/API Java/API ManufacturingModel/ API J D B C ProductModel/ API ExtendedEnterprise Strategies/API O D B C Flat Files The specifications of the prototype system. The definition of its capabilities and properties. Development of some interfaces for legacy systems and databases. Data Warehouse Knowledge, New Information Mining Models DATA MINING An Interface between Forac Experimental Platform and KD&DM Using Agent Systems Exploration of data on The Use of the Web User Accesses; Contents of Web log files; Other relevant data. Web Usage Mining Exploration of the Content of the Web Page Contents; Page links. Web Content Mining Exploration of the interconnections between hypertext documents Web Structure Mining •It represents the mining processing applied to large volumes of unstructured text. •The marketing information is available on the web as white papers, academic publications, trade journals, news, articles, reviews and even public opinions. Text mining could support the marketing professionals to efficiently use this information for finding knowledge and patterns. Semantic Web Technologies The current WWW is mainly syntactic-based where structure of the content is presented while content itself is only readable by humans. Semantic Web is directed to create and manage the future Web or at least an extention which aims to include semantics to content. Semantic Web Languages make the Web computer processable and computer understandable. The Ontology Languages are directed to formalize the Web. References • • • • • Aalst, W. and Hee, K. – Workflow Management Models, Methods, and Systems, London, New York: The MIT Press, Cooperative Information Systems, 2002. Cheun D. et al. – Advances in Knowledge Discovery and Data Mining, 5th Pacific-Asia Conference, PAKDD 2001, Hong Kong, China, Lecture Notes in AI, Berlin: Springer-Verlag, 2001. Hoover W.E. Jr. et al. – Managing the Demand-Supply Chain Value Innovations for Customer Satisfaction, New York: John Wiley & Sons, Inc. 2001. Marinescu D. – Internet-Based Workflow Management Toward a Semantic Web, New York: Wiley Series on Parallel and Distributed Computing, 2002. Schary and Skjott-Larsen – Managing the Global Supply Chain, Copenhagen: Munksgaard International Publishers Ltd., 1995. “The world is moving so fast these days that the man who says it can't be done is generally interrupted by someone doing it.” Elbert Hubbard Paul Cezanne - Foliage