Mehmet S. Aktas

Temporal representation for mining scientific data provenance

Abstract Provenance of digital scientific data is a distinct piece of metadata about a data object. It can serve as a “ground-truth” for determining the cause of execution failure for instance, or can explain a particular result to a researcher intending to reuse a data object. Provenance can quickly grow voluminous and be quite feature rich, requiring new structure and concepts that support data mining. We propose a representation of data provenance using logical time that reduces the feature space of the provenance. The temporal representation supports clustering, classification and association rule mining. This paper studies the full utility of the temporal representation through an empirical evaluation and identification of the data mining algorithms that are most effective in application to the proposed representation. The evaluation is carried out against a multi-gigabyte semi-synthetic provenance dataset built from a range of scientific workflows, and against a real one month provenance dataset gathered from a satellite instrument. Through analysis of the results via clustering metrics—purity and Normalized Mutual Information (NMI), we determine that the k -means algorithm gives the best clustering with the proposed temporal representation, while still yielding provenance-useful information.

Temporal representation for scientific data provenance

Provenance of digital scientific data is an important piece of the metadata of a data object. It can however grow voluminous quickly because the granularity level of capture can be high. It can also be quite feature rich. We propose a representation of the provenance data based on logical time that reduces the feature space. Creating time and frequency domain representations of the provenance, we apply clustering, classification and association rule mining to the abstract representations to determine the usefulness of the temporal representation. We evaluate the temporal representation using an existing 10 GB database of provenance captured from a range of scientific workflows.

BlogMiner: Web Blog Mining Application for Classification of Movie Reviews

With the increasing use of Web 2.0 platforms such as Web Blogs, discussion forums, Wikis, and various other types of social media, people began to share their experiences and opinions about products or services on the World Wide Web. Web Blogs have thus become an important source of information. In turn, great interest in blog mining has arisen, specifically due to its potential applications, such as in opinion or review search engine applications the ability to collect and analyze data. In this study, we introduce an architecture, implementation, and evaluation of a Web blog mining application, called the BlogMiner, which extracts and classifies people’s opinions and emotions (or sentiment) from the contents of weblogs about movie reviews.

Mobile Application Verification: A Systematic Mapping Study

The proliferation of mobile devices and applications has seen an unprecedented rise in recent years. Application domains of mobile systems range from personal assistants to point-of-care health informatics systems. Software development for such diverse application domains requires stringent and well-defined development process. Software testing is a type of verification that is required to achieve more reliable system. Even though, Software Engineering literature contains many research studies that address challenging issues in mobile application development, we could not have identified a comprehensive literature review study on this subject. In this paper, we present a systematic mapping of the Software Verification in the field of mobile applications. We provide definitive metrics and publications about mobile application testing, which we believe will allow fellow researchers to identify gaps and research opportunities in this field.

Structural Code Clone Detection Methodology Using Software Metrics

Unnecessary repeated codes, also known as code clones, have not been well documented and are difficult to maintain. Code clones may become an important problem in the software development cycle, since any detected error must be fixed in all occurrences. This condition significantly increases software maintenance costs and requires effort/duration for understanding the code. This research introduces a novel methodology to minimize or prevent the code cloning problem in software projects. In particular, this manuscript is focused on the detection of structural code clones, which are defined as similarity in software structure such as design patterns. Our proposed methodology provides a solution to the class-level structural code clone detection problem. We introduce a novel software architecture that provides unification of different software quality analysis tools that take measurements for software metrics for structural code clone detection. We present an empirical evaluation of our approach and investigate its practical usefulness. We conduct a user study using human judges to detect structural code clones in three different open-source software projects. We apply our methodology to the same projects and compare results. The results show that our proposed solution is able to show high consistency compared with the results reached by the human judges. The outcome of this study also indicates that a uniform structural code clone detection system can be built on top of different software quality tools, where each tool takes measurements of different object-oriented software metrics.

On the Structural Code Clone Detection Problem: A Survey and Software Metric Based Approach

Unnecessary repeated codes (clones) have not been well documented and are difficult to maintain. Code clones may become an important problem in software development cycle and they must be fixed in all occurrences. This condition increases significantly software maintenance costs and required effort/duration for understanding the code. Over the years, many techniques have been proposed in order to minimize or prevent the code cloning problems. The main focus of these techniques is on the detection of clones. In such studies, code cloning is studied under two main categories: simple and structural. Simple clone is defined as the similarity that arises from the repetition of the code snippet in the software. Structural clone is defined as the similarity in software structure (i.e. design patterns and object oriented programming class relations). Simple clone detection techniques fail to determine the reasons of code repetition whether it is due to design or not, as they do not look at the code from a wider perspective for repetitive code snippets. In this study, we survey the existing structural clones approaches. We also introduce an approach that utilizes software quality metrics for detecting the structural code clones.

An approach to hybrid personalized recommender systems

Collaborating Filtering (CF) is a recommendation method that can make predictions about a given users interest by collecting a large number of other user appreciation. Some of the major problems encountered in the use of CF are the cold-start problem and the fact that personalized recommendations cannot be done. In turn, CF-based recommendations produces ranked results where the success rate can be improved. The method proposed in this research is a hybrid recommender system that utilizes Case-Based Reasoning (CBR) in order to overcome these shortcomings and improve the success rate of the recommender system. To show the usability of the proposed hybrid recommender method, we have used a music recommendation dataset and build music listening assistant that uses the implementation of the method. The performance of the proposed method was evaluated and results are reported. The results indicate that our proposed method is successful.

Special issue on High performance computing conference (BASARIM-2017)

Numerical operations and calculations have been studied on parallel systems for many years for the purpose of achieving faster results and better performance. In recent years, these studies and their results have reached a certain level and the progress made in this regard has accelerated. The Turkish High Performance Computing Conference has been organized since 2009 for discussing challenges in high performance computing and identifying opportunities in the areas of cloud computing and big data processing for moving ahead. To this end, the integration of big data processing software stack and traditional parallel computing message passing protocols has been addressed in the keynote talk “HPC-enhanced IoT and Data-based Grid” by Geoffrey Charles Fox at the 5th Turkish High Performance Computing Conference (BASARIM-2017). With the objective of sharing and evaluating scientific research, experience, studies, and results related to high performance computing at this scientific event, this special issue presents the highlights of the program and selected high-quality papers from BASARIM-2017.

Application of provenance in social computing: A case study: Application of provenance in social computing: A case study

Complex systems such as Collective Adaptive Systems that include a variety of resources, are increasingly being designed to include people in task‐execution. Collectives, encapsulating human resources/services, represent one type of an application within which people with different type of skills can be engaged to solve one common problem or work on the same project. Mechanisms of managing social collectives are dependent on functional and non‐functional parameters of members of social collectives. In this work, we investigate the benefits provenance can offer to social computing and trade‐off implications. We show experimental results of how provenance data can help better visualize interaction and performance data during a collectives run‐time. We present novel metrics that can be derived from provenance, and lastly, we discuss privacy implications. If utilized ethically, provenance can help in developing more efficient provisioning and management mechanisms in social computing.

Implementation of Analytical Hierarchy Process in Detecting Structural Code Clones

The nature and the size of data plays an important rule at the identification process of similar objects (clones). The type of utilized similarity measures is also an important factor. The nature of data and selecting the right identification algorithm appropriate to type of data should be examined thoroughly when a clone identification technique is applied. This study suggests a new methodology in software systems for minimization/prevention of code cloning. Its main contribution is to propose an Analytical Hierarchy Process based methodology at detection of code clones in object-oriented software systems, in which the software is represented by means of software metrics data at class level. The suggested clone detection model is able to select the most suitable code clone candidates by considering different correlation and distance metrics to identify code clones. To facilitate the testing and the usability of the suggested clone detection model, the system is used for detection of structural code clone. The methodology is validated by comparison with results obtained by human judges as well as by comparison with a plain structural code clone identification approach. The evaluation of the methodology is carried out in terms of accuracy and indicates promising results.