Classifiers based on parametric or non-parametric learning methods have different advantages and disadvantages. To take advantage of the strengths of both methods, we propose an algorithm that combines a parametric mo...
详细信息
Static code analysis tools provide valuable input for experts to judge the internal quality of software. Nevertheless this approach is time consuming and therefore (semi-)automatic approaches would be desirable. In th...
详细信息
ISBN:
(纸本)9780889867154
Static code analysis tools provide valuable input for experts to judge the internal quality of software. Nevertheless this approach is time consuming and therefore (semi-)automatic approaches would be desirable. In the QBench [14] project a promising benchmarking oriented approach for calculating a quality rating was developed. We tried to apply the methods and tools of the QBench project in order to investigate the strengths and weaknesses of the approach. In summary, the calculation of the quality rating does not lead to satisfying results. We therefore developed a number of alternative calculation methods and compared them to the results of the original calculation. Some of these calculation variants lead to a better characterization of the software quality as the original QBench algorithm. Furthermore we detected some technical weaknesses in the provided toolset and considerably enhanced this toolset for calculating the quality rating to make it better applicable.
The semantic Web services composition process arranges several Web services into one composite to realize complex workflows. To do this, semantic metadata of Web services' description are used. The current approac...
The semantic Web services composition process arranges several Web services into one composite to realize complex workflows. To do this, semantic metadata of Web services' description are used. The current approaches based mainly on AI planning are immature to be used in practice. In this paper we propose an approach involving users in the semantic Web services composition to help overcome problems occurring in the composition process. The basic idea is to find the users helpful in situations when preconditions are not satisfied or some input data are not available which are in demand to create a composition.
Characteristics in the user model can be acquired and maintained in several ways, namely by means of implicit or explicit feedback, observation of user's actions, log analysis, etc. We use analysis of the content ...
详细信息
Characteristics in the user model can be acquired and maintained in several ways, namely by means of implicit or explicit feedback, observation of user's actions, log analysis, etc. We use analysis of the content that is presented to a user. We assume that comparing documents and analyzing their common and different aspects can lead to information about user's interests. We work in the Semantic Web environment where ontologies are used as a mean for content representation. We proposed a recursive method to evaluate similarity of ontological concepts. But computed similarities do not take into account user's individuality, i.e. the same results are computed for each user. User's perception of the similarity measure is subjective. Therefore, we extend the method to compute similarity with regard to the user.
Change impact analysis is a change management activity that previously has been studied much from a technical perspective. For example, much work focuses on methods for determining the impact of a change. In this pape...
详细信息
The most popular development methodologies in the last decade are based on object-oriented techniques. The goal of this paper is to investigate the possibilities of extending the object-oriented methodology of informa...
详细信息
This chapter introduces the volume on Applications of Data Mining in E-Business and Finance. It discusses how application-specific issues can affect the development of a data mining project. An overview of the chapter...
详细信息
Classifiers based on parametric or non-parametric learning methods have different advantages and disadvantages. To take advantage of the strengths of both methods, we propose an algorithm that combines a parametric mo...
详细信息
Classifiers based on parametric or non-parametric learning methods have different advantages and disadvantages. To take advantage of the strengths of both methods, we propose an algorithm that combines a parametric model (logistic regression) with a non-parametric classification method (k-nearest neighbors). This combination is based on a measure of appropriateness that uses a heuristic to decide which of the two components should contribute more to the final classification output. We measure the performance of this combination method on two data sets (one from medical informatics, and one consisting of simulated data) in terms of areas under the ROC curves (AUCs). We are able to demonstrate that our method of combining classifiers exceeds the performance of both individual classifiers taken separately.
Only a few studies exist that try to investigate whether there is a significant correlation between external software quality and the data provided by static code analysis tools. A clarification on this issue could pa...
详细信息
Only a few studies exist that try to investigate whether there is a significant correlation between external software quality and the data provided by static code analysis tools. A clarification on this issue could pave the way for more precise prediction models on the probability of defects based on the violation of programming rules. We therefore initiated a study where the defect data of selected versions of the open source development environment ldquoEclipse SDKrdquo is correlated with the data provided by the static code analysis tools PMD and FindBugs applied the source code of Eclipse. The results from this study are promising as especially some PMD rulesets show a good correlation with the defect data and could therefore serve as basis for measurement, control and prediction of software quality.
暂无评论