Catálogo de publicaciones - libros

The use of knowledge-based systems can represent an efficient approach for system management, providing automatic control strategies with Artificial Intelligence capabilities. By means of Artificial Intelligence, the system is capable of assessing, diagnosing and suggesting the best operation mode. One important Artificial Intelligence tool for automatic control is the use of fuzzy logic controllers, which are fuzzy rule-based systems comprising the expert knowledge in form of linguistic rules. These rules are usually constructed by an expert in the field of interest who can link the facts with conclusions. However, this way to work sometimes fails to obtain an optimal behavior. To solve this problem, within the framework of Machine Learning, some artificial intelligence techniques could be applied to enhance the controller behavior.

In this work, a post-processing method is used to obtain more compact and accurate fuzzy logic controllers. This method combines a new technique to perform an evolutionary lateral tuning of the linguistic variables with a simple technique for rule selection (that removes unnecessary rules). To do so, the tuning technique considers a new rule representation scheme by using the linguistic 2-tuples representation model which allows the lateral variation of the involved linguistic labels.

The path-based indexing methods such as the three-dimensional bitmap indexing have been used for collecting and retrieving the similar XML documents. To do this, the paths become the fundamental unit for constructing index. In case the document structure changes, the path extracted before the change and the one after the change are regarded as totally different ones regardless of the degree of the change. Due to this, the performance of the path-based indexing methods is usually bad in retrieving and clustering the documents which are similar. A novel method which can detect the similar paths is needed for the effective collecting and retrieval of XML documents. In this paper, a new path construction similarity which calculates the similarity between the paths is defined and a path bitmap indexing method is proposed to effectively load and extract the similar paths. The proposed method extracts the representative path from the paths which are similar. The paths are clustered using this, and the XML documents are also clustered using the clustered paths. This solves the problem of existing three-dimensional bitmap indexing. Through the performance evaluation, the proposed method showed better clustering accuracy over existing methods.

In this paper we propose an algorithm, called the modified suppressed fuzzy c-means (MS-FCM), that simultaneously performs clustering and parameter selection for the suppressed FCM (S-FCM) proposed by Fan et al. [2]. Numerical examples illustrate the effectiveness of the proposed MS-FCM algorithm. Finally, the S-FCM and MS-FCM algorithms are applied in the segmentation of the magnetic resonance image (MRI) of an ophthalmic patient. In our comparisons of S-FCM, MS-FCM and alternative FCM (AFCM) proposed by Wu and Yang [14] for these MRI segmentation results, we find that the MS-FCM provides better detection of abnormal tissue than S-FCM and AFCM when based on a window selection. Overall, the MS-FCM clustering algorithm is more efficient and is strongly recommended as an MRI segmentation technique.

This paper presents two new clustering algorithms which are based on the entropy regularized fuzzy -means and can treat data with some errors. First, the tolerance which means the permissible range of the error is introduced into optimization problems which relate with clustering, and the tolerance is formulated. Next, the problems are solved using Kuhn-Tucker conditions. Last, the algorithms are constructed based on the results of solving the problems.

Hard -means can be used for building classifiers in supervised machine learning. For example, in a -class problem, clusters are built for each of the classes. This results into . centroids. Then, new examples can be classified according to the nearest centroid.

In this work we consider the problem of building classifiers using fuzzy clustering techniques. In particular, we consider the use of fuzzy -means, as well as some variations. Namely, fuzzy -means with variable size and entropy based fuzzy -means.