Nasraoui, O. and R. Krishnapuram, 2000. A novel approach to unsupervised robust clustering using genetic niching. In Ninth IEEE Intl. Conf. Fuzzy Systems, San Antonio, TX, pp: 170-175.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....patterns in Fig. 1(a) The aim of a CA algorithm is to automatically identify the two clusters present, as shown in Fig. 1(b) A good algorithm can also autonomously recognize the number of the clusters, even if it is not known apriori and also in the presence of noise outlier points. [2, 4, 23 26]. While, if we apply a VQ technique to the same data set, the number of cells into which the data have to be subdivided depends only on the desired degree of approximation. The higher the number of the codewords (and likewise of the cells) the better the degree of approximation. 4 and 8 cells ....

O. Nasraoui and R. Krishnapuram, \A novel approach to unsupervised robust clustering using genetic niching," in Proceedings of the Ninth International Conference on Fuzzy Systems, (San Antonio, TX), May 2000.

No context found.

Nasraoui, O. and R. Krishnapuram, 2000. A novel approach to unsupervised robust clustering using genetic niching. In Ninth IEEE Intl. Conf. Fuzzy Systems, San Antonio, TX, pp: 170-175.

No context found.

O. Nasraoui and R. Krishnapuram, "A novel approach to unsupervised robust clustering using genetic niching," in In Proceedings of the Ninth IEEE International Conference on Fuzzy Systems, pp. 170--175, 2000.

No context found.

O. Nasraoui and R. Krishnapuram, "A novel approach to unsupervised robust clustering using genetic niching," in In Proceedings of the Ninth IEEE International Conference on Fuzzy Systems, pp. 170--175, 2000.

No context found.

O. Nasraoui, and R. Krishnapuram, A Novel Approach to Unsupervised Robust Clustering using Genetic Niching, Proc. of the 9

....raised serious privacy concerns, are subjective, and do not adapt to changing interests. Mass profiling, on the other hand, is based on general trends of usage (thus protecting privacy) and can be achieved by mining or discovering user profiles from historical data stored in server access logs [1,2,3,4,5,6] . A complete intelligent Web personalization system is generally based on Web usage mining to discover useful knowledge about user access patterns, followed by a recommendation system to act on this knowledge in order to respond to the users individual interest, in a manner transparent to the ....

....relative to a group of users sharing similar interest . This work is supported by a National Science Foundation CAREER Award (NSF IIS 0133948) to O. Nasraoui. User profile, group of users, and interest similarity are concepts for which we have previously provided abstract and computational models [3,4,5,6]. Hence, our framework for Web mining consists of models for clusters (to model the groups) Web or user sessions (to model users) a Web session similarity measure (to capture similar interests) and user profiles (to capture user interests) be prone to significant amounts of error and ....

[Article contains additional citation context not shown here]

O. Nasraoui, and R. Krishnapuram, "A Novel Approach to Unsupervised Robust Clustering using Genetic Niching," Proc. of the 9th IEEE International Conf. on Fuzzy Systems, San Antonio, TX, May 2000, pp. 170-175.

....and others based on modifying the objective of the fuzzy centroid mean to make the parameter estimate more robust to noise [14, 2, 8] In all of these methods, the number of cluster is given in advance. When the number of clusters is not known, the clustering techniques are called unsupervised [13, 9, 18]. We propose a novel unsupervised robust clustering technique based on the Gravitational Law and the sec University of Memphis Department of Mathematical Sciences, Computer Sciences Division and Universidad Nacional de Colombia IThe University of Memphis Department of Mathematical Sciences, ....

O. Nasraoui and R. Krishnapuram, A Novel Approach to Unsupervised Robust Clustering Using Genetic Niching. In Proceedings of the Ninth IEEE International Conference on Fuzzy Systems, 170-175, 2000.

No context found.

O. Nasraoui and R. Krishnapuram, A novel approach to unsupervised robust clustering using genetic niching, 9th IEEE Int. Conf. Fuzzy Syst., San Antonio, TX, May 2000, 170--175.

....generations) than a purely genetic search . When mating takes place, each child should inherit the scale parameter, of the closest parent as its initial scale before updating. After convergence of the population, the best individual from each good niche is extracted using a greedy approach [21] to obtain the set of final cluster centers, 3.1.3 Computational Complexity In each generation, the most extensive computational requirement for UNC consists of computing the residuals, fitness and scale, for each of the = individuals in the population, and exactly = inter niche ....

....to a certain degree in . The cosine based dissimilarity measure, defined in Section 2.2, is used to compute the distance between session data and candidate profiles. The formulation of our fitness function is based on a density based measure that was proposed for mining clusters in [21]. The fitness value, for the candidate profile, is defined as the density of a hypothetical cluster of Web sessions with as a summarizing prototype or medoid. It is defined as 5 weight that measures how typical a session profile, and is given by H, H A ....

O. Nasraoui and R. Krishnapuram, "A novel approach to unsupervised robust clustering using genetic niching," in Ninth IEEE International Conference on Fuzzy Systems, San Antonio, TX, May 2000, pp. 170--175.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC