Abstract In this paper, a new approach for fitness evaluation of Genetic Algorithm (GA) is discussed. The approach is based on position of the genes in a chromosome. A random initial population sometimes may generate recessive strings that may increase the number of iterations. Also, it is hard to decide time to arrive at a particular solution. Chromosome’s strength is dependent... In this paper, a new approach for fitness evaluation of Genetic Algorithm (GA) is discussed. The approach is based on position of the genes in a chromosome. A random initial population sometimes may generate recessive strings that may increase the number of iterations. Also, it is hard to decide time to arrive at a particular solution. Chromosome’s strength is dependent on the arrangement of its genes. A new position based fitness evaluation technique is introduced, where fitness is assigned to each gene’s position to compute the strength of a chromosome. Also selecting the weakest position for crossing over avoids the probability of dummy iterations and increases the efficiency. We have tested the new approach on “N” Queen Problem that represents NP hard problems. The experimental results are comparatively good with other state of art GAs, simple genetic algorithm (SGA), enhanced improved genetic algorithm (EIGA) and adaptive genetic algorithm (AGA). The strength of new fitness approach increases in fitness in each iteration and efficient. Potential application includes search techniques and machine learning. Read More Read Less

Abstract This work proposes a study of algorithms used for attribute selection in text clusterization in the scientific literature and selection of attribute in the Cassiopeia model The most relevant contributions include the use of summarized texts as an entrance in pre-processing stage of clusterization, language independence with the use of stop words and the treatment of high dimensionality, a problem... This work proposes a study of algorithms used for attribute selection in text clusterization in the scientific literature and selection of attribute in the Cassiopeia model The most relevant contributions include the use of summarized texts as an entrance in pre-processing stage of clusterization, language independence with the use of stop words and the treatment of high dimensionality, a problem that is inherent to Text Mining. Hence, our intention is to achieve an improvement in the measurement of clusters as well as to solve the problem of high dimensionality. Read More Read Less

Abstract As computer systems become more complex, the state of their inner workings become more and more important to the system administrators working to keep them running. Log files provide much needed visibility into these systems, whether they are hardware, operating systems or applications. Unfortunately, systems can easily create overwhelming amounts of data for administrators to comb through. This research tests... As computer systems become more complex, the state of their inner workings become more and more important to the system administrators working to keep them running. Log files provide much needed visibility into these systems, whether they are hardware, operating systems or applications. Unfortunately, systems can easily create overwhelming amounts of data for administrators to comb through. This research tests the effectiveness of three off-the-shelf Bayesian spam email filters (SpamAssassin, SpamBayes and Bogofilter) for effectiveness as log entry classifiers. A simple scoring system, the Filter Effectiveness Scale (FES), is proposed and used to compare these three filters. The filters are tested in three stages: 1) the filters were tested with the SpamAssassin corpus, with various manipulations made to the messages, 2) the filters were tested for their ability to differentiate log entries from two services, with logs taken from production systems, and 3) the filters were trained with problem-related log entries from system outages and then tested for effectiveness in finding similar outages through similar log files. Read More Read Less

Abstract Much of the recent research has been aimed at improving web performance and scalability. For attaining the goal of improving web performance the basic need is the understanding of WWW work loads. In this paper we present a method useful for the system engineer to improve the service performance of a Web server through session-based Web workload, the best indicator... Much of the recent research has been aimed at improving web performance and scalability. For attaining the goal of improving web performance the basic need is the understanding of WWW work loads. In this paper we present a method useful for the system engineer to improve the service performance of a Web server through session-based Web workload, the best indicator of the users perception of the Web quality. Bytes transferred per session is one of the characteristics of intra-session which collectively describe session-based Web workload. This characteristic exhibits heavy-tailed behavior and its distribution match well with the Pareto Type I distribution [Goseva-Popstojanova et al. (2006)]. So for the performance study, we estimate the probability, R = P(X > Y ), when X and Y are two independent but not identically distributed random variables following Pareto Type I distribution, using the maximum likelihood estimator and Hill estimator. Extensive simulation studies are carried out to study the performance of these estimators. A generalized two-sided confidence interval for R of the Pareto type I distribution is constructed. The derived confidence interval suits both small samples and large samples. The average width and the coverage probability of this confidence interval is compared with the usual asymptotic confidence intervals through simulations. Using real data, we illustrate how R and generalized confidence interval of R can be used for improving the service performance of a Web server. Read More Read Less