A fuzzy c-means variant for the generation of fuzzy term sets

Liao, T. Warren; Celmiņš, Aivars; Hammell, Robert J.

doi:10.1016/s0165-0114(02)00136-7

Cited by 73 publications

(27 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This process is usually performed by non supervised clustering techniques (López, Magdalena & Velasco, 1999;Liao, Celmins & Hammell II, 2003), data equalization techniques (Pedrycz, 2001) or by an ascending method based on fuzzy set merging (Guillaume & Charnomordic, 2003). These approaches consider only information related to input variables, and no assumption about the output is made.…”

Section: Fuzzy Partition Designmentioning

confidence: 99%

“…In addition, our interpretability quest requires some specific properties for the partitions. The result is that those clustering techniques generating multidimensional clusters (Liao, Celmins & Hammell II, 2003) cannot be applied, since we need one-dimensional membership functions, obtained by independently partitioning the universe of each variable. As a general approach, it is possible to use any one-dimensional optimization technique if it includes some semantic constraints.…”

Section: Fuzzy Partition Designmentioning

confidence: 99%

See 1 more Smart Citation

Expert guided integration of induced knowledge into a fuzzy knowledge base

Guillaume¹,

Magdalena

2006

Soft Comput

View full text Add to dashboard Cite

This paper proposes a method for building accurate and interpretable systems by integrating expert and induced knowledge into a single knowledge base. To favor the cooperation between expert knowledge and data, the induction process is run under severe constraints to ensure the fully control of the expert. The procedure is made up of two hierarchical steps. Firstly, a common fuzzy input space is designed according to both the data and expert knowledge. The compatibility of the two types of partitions, expert and induced, is checked according to three criteria: range, granularity and semantic interpretation. Secondly, expert rules and induced rules are generated according to the previous common fuzzy input space. Then, induced and expert rules have to be merged into a new rule base. Thanks to the common universe resulting from the first step, rule comparison can be made at the linguistic level only. The possible conflict situations are managed and the most important rule base features, consistency, redundancy and completeness, are studied. The first step is thoroughly described in this paper, while the second is only introduced.

show abstract

Section: Fuzzy Partition Designmentioning

confidence: 99%

Section: Fuzzy Partition Designmentioning

confidence: 99%

Expert guided integration of induced knowledge into a fuzzy knowledge base

Guillaume¹,

Magdalena

2006

Soft Comput

View full text Add to dashboard Cite

show abstract

“…The rule based fuzzy systems (RBFS), focus of interest for this work, usually have two main components: a knowledge base (KB) and an inference mechanism (IM). There are several approaches for the automatic definition of data bases and automatic generation of knowledge bases from data, including clustering algorithms, neural networks, and Genetic Algorithms (GA), which are among the well-succeeded ones [18][19][20][21][22]. There has been a considerable research effort focussed on the use of GA in the design of fuzzy systems.…”

Section: Introductionmentioning

confidence: 99%

Fuzzy Features Selection Technique for Brain MR Images

Meenakshi¹,

Dubey²

2013

IJCA

View full text Add to dashboard Cite

Brain tumors are the second leading cause of cancer deaths in human throughout the world. Therefore accurate diagnosis is important for successful treatment of brain tumor. When modeling a complex, poorly defined problem with hundreds of possible inputs one must identify the significant inputs before any known modeling techniques can be applied. As generally the data contains many redundant features in it. Redundant features are those which provide no more information than the selected features and irrelevant features provide no more useful information in any context. Feature selection is the process in which one can select a subset of relevant features for use in model construction. Features selection techniques are very useful because they provide: improved model interpretability, shorter training times and enhanced generalization by reducing over fitting. For offering these benefits features selection techniques have become an apparent need in many bio-informatics applications such as brain MR images. In order to get best information a new fuzzy feature selection approach for MR brain tumor image have been used for the first time. The proposed method presents the two stages fuzzy curves and fuzzy surfaces. Fuzzy curves are used to isolate significant features from input and fuzzy surfaces to eliminate input features dependent on the significant input features. Results show that this method is able to achieve a very good reduction in the number of features.

show abstract

“…Chen and Wang 5 presented an enhanced FCM algorithm for initializing a fuzzy model, which incorporates a fuzzy validity function to find the optimal number of clusters, c, and a heuristic method to calibrate the fuzzy exponent, m. Runkler and Bezdek 25 developed an approach to obtain the LHS membership functions of a smooth first-order Takagi Sugeno system, given the points and the slopes of the RHS functions. Most recently, Liao, et al 15 developed a fuzzy c-means variant for the generation of an optimal set of fuzzy terms for each feature treated as a linguistic variable.…”

Section: Introductionmentioning

confidence: 99%

A fuzzy c-medians variant for the generation of fuzzy term sets

Liao

2002

Int. J. Intell. Syst.

View full text Add to dashboard Cite

A fuzzy c-medians variant is proposed for the generation of fuzzy term sets with one half overlap. The proposed variant is modified from the original algorithm mainly in two areas. The first modification ensures that two end terms take the maximum and minimum domain values as their centers. The second modification prevents the generation of non-convex fuzzy terms that often are obtained by the original algorithm. The optimal number of terms and the optimal shape of membership function associated with each term are determined based on the mean squared error criterion. The exponential weight, m, used in the algorithm is found to greatly affect the shape of membership function. The effect of data size used for the generation of fuzzy terms is also discussed. A generalized s-shaped function with a tunable parameter along with its complement (i.e., a generalized z-shaped function with a tunable parameter) is developed to fit all term sets generated by the fuzzy c-medians variant using various m values.

show abstract

A fuzzy c-means variant for the generation of fuzzy term sets

Cited by 73 publications

References 27 publications

Expert guided integration of induced knowledge into a fuzzy knowledge base

Expert guided integration of induced knowledge into a fuzzy knowledge base

Fuzzy Features Selection Technique for Brain MR Images

A fuzzy c-medians variant for the generation of fuzzy term sets

Contact Info

Product

Resources

About