Search

Most of the popular Big Data analytics tools evolved to adapt their working environment to extract valuable information from a vast amount of unstructured data. The ability of data mining techniques to filter this helpful information from Big Data led to the term Big Data Mining. Shifting the scope of data from small-size, structured, and stable data to huge volume, unstructured, and quickly changing data brings many data management challenges. Different tools cope with these challenges in their own way due to their architectural limitations. There are numerous parameters to take into consideration when choosing the right data management framework based on the task at hand. In this paper, we present a comprehensive benchmark for two widely used Big Data analytics tools, namely Apache Spark and Hadoop MapReduce, on a common data mining task, i.e., classification. We employ several evaluation metrics to compare the performance of the benchmarked frameworks, such as execution time, accuracy, and scalability. These metrics are specialized to measure the performance for classification task. To the best of our knowledge, there is no previous study in the literature that employs all these metrics while taking into consideration task-specific concerns. We show that Spark is 5 times faster than MapReduce on training the model. Nevertheless, the performance of Spark degrades when the input workload gets larger. Scaling the environment by additional clusters significantly improves the performance of Spark. However, similar enhancement is not observed in Hadoop. Machine learning utility of MapReduce tend to have better accuracy scores than that of Spark, like around 3%, even in small size data sets., Comment: 2021 5th International Conference on Cloud and Big Data Computing (ICCBDC 2021)

The metabolic wiring of patient cells is altered drastically in many diseases, including cancer. Understanding the nature of such changes may pave the way for new therapeutic opportunities, as well as the development of personalized treatment strategies for patients. In this paper, we propose an algorithm, Metabolitics, which allows systems-level analysis of changes in the biochemical network of cells in disease states. It enables the study of a disease at both reaction- and pathway-level granularities for detailed and summarized view of disease etiology. Metabolitics employs flux variability analysis with an objective function which is dynamically built based on the biofluid metabolomics measurements in a personalized manner. Moreover, Metabolitics builds classification models for each disease to diagnose patients and predict their subgroups based on the computed metabolic network changes. We demonstrate the use of Metabolitics on three distinct diseases, namely, breast cancer, Crohn's disease, and colorectal cancer. Our results show that the constructed supervised learning models successfully diagnose patients by f1-score of over 90% on the average. Besides, in addition to the confirmation of previously reported breast cancer associated pathways, we discovered that Butanoate metabolism experiences significantly decreased activity, while Arginine and Proline Metabolism is subject to significant increase in activity, which have not been reported before. Metabolitics is made available as a Python package in pypi.

With the dramatic increase in the amount of the text-based data which commonly contains misspellings and other errors, querying such data with flexible search patterns becomes more and more commonplace. Relational databases support the LIKE operator to allow searching with a particular wildcard predicate (e.g., LIKE 'Sub%', which matches all strings starting with 'Sub'). Due to the large size of text data, executing such queries in the most optimal way is quite critical for database performance. While building the most efficient execution plan for a LIKE query, the query optimizer requires the selectivity estimate for the flexible pattern-based query predicate. Recently, SPH algorithm is proposed which employs a sequence pattern-based histogram structure to estimate the selectivity of LIKE queries. A drawback of the SPH approach is that it often overestimates the selectivity of queries. In order to alleviate the overestimation problem, in this paper, we propose a novel sequence pattern type, called positional sequence patterns. The proposed patterns differentiate between sequence item pairs that appear next to each other in all pattern occurrences from those that may have other items between them. Besides, we employ redundant pattern elimination based on pattern information content during histogram construction. Finally, we propose a partitioning-based matching scheme during the selectivity estimation. The experimental results on a real dataset from DBLP show that the proposed approach outperforms the state of the art by around 20% improvement in error rates.

In this paper, new representations of a Bertrand curve pair in three dimensional Lie groups with bi-invariant metric are given. Besides, the spherical indicatrices of a Bertrand curve pair are obtain and the relations between the spherical indicatrices and new representations of Bertrand curve pair are shown., Comment: 12 pages

Let $M$ be an $n-$dimensional differentiable manifold with a symmetric connection $\nabla $ and $T^{\ast}M$ be its cotangent bundle. In this paper, we study some properties of the modified Riemannian extension $% \widetilde{g}_{\nabla,c}$ on $T^{\ast}M$ defined by means of a symmetric $% (0,2)$-tensor field $c$ on $M.$ We get the conditions under which $T^{\ast}M $ endowed with the horizontal lift $^{H}J$ of an almost complex structure $J$ and with the metric $\widetilde{g}_{\nabla,c}$ is a K\"{a}hler-Norden manifold. Also curvature properties of the Levi-Civita connection and another metric connection of the metric $\widetilde{g}_{\nabla,c}$ are presented.