Catálogo de publicaciones - libros

Fuzzy Multiple Criteria Decision Making (MCDM) has been widely used in evaluating and ranking weapon systems characterized by fuzzy assessments with respect to multiple criteria. However, most criteria have interdependent or interactive characteristic so weapon system cannot be evaluated by conventional evaluation methods. In this paper, a new method based on trapezoidal fuzzy AHP and fuzzy Integral is proposed. The ratings of criteria performance are described by linguistic terms expressed in trapezoidal fuzzy numbers. The weights of the criteria are obtained by trapezoidal fuzzy AHP. And the hierarchical fuzzy integral model is proposed based on -fuzzy measure and Sugeno integral to determine the synthesis evaluation of weapon system. Finally, an example of evaluating the best main battle tank is given. The results demonstrate the engineering practicability and effectiveness of this method.

This paper presents a new approach for the cost-sensitive classification problems based on the Boosting ensemble of support vector machines (SVMs). Different from conventional Boosting ensemble learning methods that adjust the distribution of training instances for minimizing the misclassification rate, the presented approach adjusts the training data distribution so as to minimize the expected cost of classification. This approach has been applied successfully in Microcalcification (MC) detection which is a typical cost-sensitive classification problem in breast cancer diagnosis. Its performance is evaluated by means of Receiver Operating Characteristics (ROC) curves and the expected costs of classification. Experimental results have consistently confirmed that the ROC of the SVM ensemble classifier is very close to the curve enveloping the base classifier ROC curves. This characteristic illustrates that the SVM ensemble is able to always improve the performance of the classification. Furthermore, the experimental results demonstrate that the method presented is able to not only increase the area under the ROC curve (AUC) but also minimize the expected classification cost.

Let opinions of experts among group decision making be represented as L-R fuzzy numbers. The difference of two experts’ opinions is reflected by two distances, which are called the left-hand side distance and the right-hand side one. A method to calculate two types of distances based on the same -level is presented. Then the distances are employed to construct a new similarity function to measure the similarity degrees of both sides which represent the pessimistic and optimistic similarity degrees between the experts, respectively. The degree of importance of each expert among group decision making is obtained by employing Saaty’s analytic hierarchy process (AHP). The method of aggregating individual fuzzy opinions into a group consensus opinion by combining similarity degrees and the degree of importance of each expert is proposed. Finally some properties of the proposed similarity measure are proved and some numeric examples are shown to illustrate our method.

To enhance resolution and reduce artifacts in imaging from incomplete projections, a novel imaging model and algorithm to imaging from incomplete projections—multi-criterion fuzzy optimization approach is presented. This model combines fuzzy theory and multi-criterion optimization approach. The membership function is used to substitute objective function and the minimum operator is taken as fuzzy operator. And a novel resolution method was proposed. The result reconstructed from computer-generated noisy projection data is shown. Comparison of the reconstructed images indicates that this algorithm gives better results both in resolution and smoothness over analytic imaging algorithm and conventional iterative imaging algorithm.

Mining frequent item sets (frequent patterns) in transaction databases is a well known problem in data mining research. This work proposes a sampling-based method to find frequent patterns. The proposed method contains three phases. In the first phase, we draw a small sample of data to estimate the set of frequent patterns, denoted as . The second phase computes the actual supports of the patterns in as well as identifies a subset of patterns in that need to be further examined in the next phase. Finally, the third phase explores this set and finds all missing frequent patterns. The empirical results show that our algorithm is efficient, about two or three times faster than the well-known FP-growth algorithm.

In this paper, first, the model of a fuzzy reversed posynomial geometric programming is built after introduction of a prime fuzzy posynomial geometric programming. Besides, its fuzzy Lagrange problem is studied. On this basis, a validity of direct algorithm is designed to solve the former programming. Finally, the model and algorithm are testified by a numerical example.

Documents that are widely in use today included many atypical characteristics. In particular, non-standardization appears more frequently in e-mail documents than other documents due to the extensive use of informal expressions such as slang and abbreviation. Automatic document classification may differ significantly according to the characteristics of documents that are subject to classification, as well as classifier’s performance. We suggest a three-step preprocessing algorithm by stages for accurate automatic classification for each e-mail category. This research identifies e-mail document’s characteristics to apply a three-step preprocessing algorithm that can minimize e-mail document’s atypical characteristics.

This paper presents an analysis of product quality improvement in ultra-precision manufacturing industry using data mining for developing quality improvement strategies. Based on 11320 ultra-precision optical products that were produced from the study factory during the period of June 1 and August 31, 2004, important factors impacting the product quality were identified via the decision tree method for data mining. Findings showed that the important factors for the percentage of defectives were type of processing chain, precision requirement, product classes, and raw material. The optimum range of target group in production quality indicators was identified from the gains chart.

Intrusion detection is a critical component of secure information system. Recently applying artificial intelligence, machine learning and data mining techniques to intrusion detection system are increasing. But most of researches are focused on improving the classification performance of classifier. Selecting important features from input data lead to a simplification of the problem, faster and more accurate detection rates. Thus selecting important features is an important issue in intrusion detection. Another issue in intrusion detection is that most of the intrusion detection systems are performed by off-line and it is not proper method for realtime intrusion detection system. In this paper, we develop the realtime intrusion detection system which combining on-line feature extraction method with Least Squares Support Vector Machine classifier. Applying proposed system to KDD CUP 99 data, experimental results show that it have remarkable feature feature extraction and classification performance compared to existing off-line intrusion detection system.

We proposed a new algorithm SuffixMiner which eliminates the requirement of multiple passes through the data when finding out all frequent itemsets in data streams, takes full advantage of the special property of suffix-tree to avoid generating candidate itemsets and traversing each suffix-tree during the itemset growth, and utilizes a new itemset growth method to mine all frequent itemsets in data streams. Experiment results show that the SuffixMiner algorithm not only has an excellent scalability to mine frequent itemsets over data streams, but also outperforms Apriori and Fp-Growth algorithms.