- Title
- A smart experience-based knowledge analysis system (SEKAS)
- Creator
- Wang, Peng
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2014
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- This thesis addresses issues associated with using ever-increasing amounts of information and knowledge more effectively, and taking advantage of knowledge generated through experience. With very fast expansion of the Internet has created several problems and challenges linked to the increasing amount of information in Web content. These challenges are related mainly to the difficulty of extracting potentially useful information and knowledge from Internet pages. Data mining is a tool that enables enterprises to learn from existing experience by providing them with useful and accurate trends about their customers’ behaviour, and assists organisations in predicting which products their customers may be interested in buying. Moreover, in the real world, it is common to face optimisation problems that have two or more objectives that must be optimised at the same time, that are typically explained in different units, and are in conflict with one another. The evolutionary algorithm can use experience that is derived from a former decision event to improve the evolutionary algorithm’s ability to find optimisation solutions rapidly and efficiently. A hybrid structure, the Smart Experience-based Knowledge Analysis System (SEKAS), is put forward in this thesis to address issues of knowledge management and use. SEKAS combines a set of experience knowledge structures (SOEKS) with multiple techniques to provide a comprehensive knowledge management approach capturing, discovering, reusing and storing knowledge for the users. The SEKAS integrates a novel Decisional DNA (DDNA) knowledge structure with the traditional web crawler technologies. DDNA, as a knowledge representation platform, can help deal with noisy and incomplete data, with learning from experience, and with making precise decisions and predictions in vague and fuzzy environments. This thesis outlines the investigation of the combination of DDNA and feature selection algorithms to guarantee the future performance for prediction. The proposed approaches are general and extensible in terms of both designing novel algorithms, and in the application to other domains. The SEKAS integrates the evolutionary algorithm, Non-dominated Sorting Genetic Algorithm II (NSGA-II) (Deb et al. 2002), using experience that is derived from a former decision event, to improve the evolutionary algorithm’s ability to find optimal solutions rapidly and efficiently. The SEKAS application to solve a travelling salesman's problem shows that this new proposed hybrid model can find optimal, or close to true, Pareto-optimal solutions in a fast and efficient way. Several conceptual elements for this thesis have been implemented in the testing prototype, and the experimental results that were obtained show that the SEKAS system has great potential for managing knowledge, as well as improving the response times for providing accurate solutions. Consequently, the SEKAS can provide a universal knowledge management platform for mass autonomous mechanisms and provides many functionalities for improving the efficiency in the organisational decision-making process. A real-world implementation in clinical domain is also provided in this thesis. Clinical decisional events are acquired and formalised inside the system by using the experiential knowledge representation techniques SOEKS and Decisional DNA. Three different algorithms are then applied to the clinical experience, to provide a weighting of the different decision criteria, their fine-tuning, and the formalisation of new ones.
- Subject
- knowledge management; evolutionary algorithm; heuristic
- Identifier
- http://hdl.handle.net/1959.13/1054153
- Identifier
- uon:15711
- Rights
- Copyright 2014 Peng Wang
- Language
- eng
- Full Text
- Hits: 534
- Visitors: 1029
- Downloads: 307
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Abstract | 642 KB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Thesis | 4 MB | Adobe Acrobat PDF | View Details Download |