The case-based reasoning (CBR) becomes a novel paradigm that solves a new problem by remembering a previous similar situation and by reusing information and knowledge of that situation. However, the acquisition of case knowledge is a bottleneck within case-based reasoning. The use of rough set and data mining to discover knowledge from traditional database and to construct case base is desired. In this paper we discuss, in detail, the approach taken to acquire case knowledge. Rough set is used to preprocess the raw data that is noisy and redundant on the attribute. A Kohonen network is proposed to identify initial clusters within the data having been preprocessed. These clusters are then analyzed using C.45 and non-unique clusters are grouped to form concepts. Cases are then chosen from each of the identified concepts as well as outliers in the database. The results indicate that the proposed approach achieves a high reduction in the size of the case base.
[1]
Theresa Beaubouef,et al.
Rough set techiques for uncertainty management in automated story generation
,
1998,
ACM-SE 36.
[2]
Qiang Yang,et al.
Redundancy Detection in Semistructured Case Bases
,
2001,
IEEE Trans. Knowl. Data Eng..
[3]
C. J. V. Rijsbergen,et al.
Rough Sets, Fuzzy Sets and Knowledge Discovery
,
1994,
Workshops in Computing.
[4]
D. Massart,et al.
Application of rough set theory to feature selection for unsupervised clustering
,
2002
.
[5]
Janusz Zalewski,et al.
Rough sets: Theoretical aspects of reasoning about data
,
1996
.
[6]
Miltos Petridis,et al.
Intelligent design assistant (IDA): a case base reasoning system for material and design
,
2001
.
[7]
Ni Zhi.
Automatic Case Acquisition in Case-Based System Using Data Mining
,
2003
.