Search

The analysis and use of big data (BD) in today's enterprise context is dominated by high data quality (DQ). Decisions and opinions derived from poor quality data have negative and unexpected consequences for organizations. Currently, in the data quality field, researchers urgently need to consider improving the big data quality (BDQ), especially because of the lack of detailed research in this area. Considering the big data quality will allow organizations to take sensible, accurate assessment-based decisions. This article examines and summarizes current research into big data quality by discovering the fundamental qualities of big data and its quality. In big data, the main issues of data quality are also examined. The suggestions of some researchers on data quality evaluation are outlined. Examining this subject can help to improve conceptual measurements of high data quality and provide a strong basis for future researches by developing integrated analysis and evaluation of data quality models using appropriate algorithms

Automatic themes-based classification of Quran verses is the process of classifying verses to predefined categorizes or themes. It is an essential task for all Muslims and people interested in studying the Quran. Quran themes-based classification could be used in many natural language processing (NLP) fields such as search engines, data mining, question–answering systems, and information retrieval applications. This paper presents an ensemble multi-label classification model that automatically identifies and classifies the Quran verses based on themes/topics. The model is composed of four phases: pre-processing, data vectorization, binary relevance classifier, and voting module. Firstly, the verses of the second chapter of the Quran (Al-Baqarah) are tokenized and normalized. Then, the topics of these verses are manually labeled based on “Mushaf Al-Tajweed” classification. Secondly, verses are converted into features’ vectors using term frequency-inverse document frequency (TF-IDF) and word2vec techniques. Word2vec is used to consider the semantic meaning of Quranic words and to improve performance. Also, they are trained on a collected classic Arabic corpus of 200 million words. Then, the binary relevance multi-label classification technique is applied using three different classifiers: logistic regression, support vector machine, and random forest, which categorize verses into 393 topics/tags. Finally, the voting module is applied by picking the tags with the maximum prediction probability among the three classifiers. The results of the three classifiers and the ensemble model are compared against “Mushaf Al-Tajweed.” The ensemble model outperforms the three classifiers. Its average hamming loss, recall, precision, and F1-Score are 0.224, 81%, 75%, and 77%, respectively.

Leukocytes, or white blood cells (WBCs), are microscopic organisms that fight against infectious disease, bacteria, viruses, and others. The manual method to classify and count WBCs is tedious, time-consuming and may has inaccurate results, whereas the automated methods are costly. The objective of this work is to automatically identify and classify WBCs in a microscopic image into four types with higher accuracy. BCCD is the used dataset in this study, which is a scaled down blood cell detection dataset. BCCD is firstly pre-processed by passing through several processes such as segmentation and augmentation,then it is passed to the proposed model. Our model combines the privilege of deep models in automatically extracting features with the higher classification accuracy of traditional machine learning classifiers.The proposed model consists of two main layers; a shallow tuning pre-trained model and a traditional machine learning classifier on top of it. Here, ten different pretrained models with six different machine learning are used in this study. Moreover, the fully connected network (FCN) of pretrained models is used as a baseline classifier for comparison. The evaluation process shows that the hybrid between MobileNet-224 as feature extractor with logistic regression as classifier has a higher rank-1 accuracy with 97.03%. Besides, the proposed hybrid model outperformed the baseline FCN with 25.78% on average.

Background:The p7-transactivated protein1 of Hepatitis C virus is a small integral membrane protein of 127 amino acids, which is crucial for assembly and release of infectious virions. Ab initio or comparative modelling, is an essential tool to solve the problem of protein structure prediction and to comprehend the physicochemical fundamental of how proteins fold in nature.Results:Only one domain (1-127) of p7-transactivated protein1 has been predicted using the systematic in silico approach, ThreaDom. I-TASSER was ranked as the best server for full-length 3-D protein structural predictions of p7-transactivated protein1 where the benchmarked scoring system such as C-score, TM-score, RMSD and Z-score are used to obtain quantitative assessments of the I-TASSER models. Scanning protein motif databases, along with secondary and surface accessibility predictions integrated with post translational modification sites (PTMs) prediction revealed functional and protein binding motifs. Three protein binding motifs (two Asp/Glutamnse, CTNNB1- bd_N) with high sequence conservation and two PTMs prediction: Camp_phospho_site and Myristyl site were predicted using BLOCKS and PROSITE scan. These motifs and PTMs were related to the function of p7-transactivated protein1 protein in inducing ion channel/pore and release of infectious virions. Using SCOP, only one hit matched protein sequence at 71-120 was classified as small proteins and FYVE/PHD zinc finger superfamily.Conclusion:Integrating this information about the p7-transactivated protein1 with SCOP and CATH annotations of the templates facilitates the assignment of structure–function/ evolution relationships to the known and the newly determined protein structures.

Decomposition based Multi Objectives Evolutionary Algorithms (MOEA/D) became one of research focus in the last decade. That is due to its simplicity as well as its effectiveness in solving both constrained and unconstrained problems with different Pareto Front (PF) geometries. This paper presents the challenges on different research areas concerning MOEA/D. Firstly, the original MOEA/D algorithm is explained. Its basic idea is to decompose the Multi Objectives Optimization (MOO) problem into multiple single objective optimization sub problems and works concurrently to solve these sub problems. Each sub problem is optimized with the help of the information gained from its neighborhood. Then, two major factors that affect the search ability of decomposition based MOO algorithms: Scalarization Functions (SF) and weight vectors generation and adaptation are discussed. Furthermore, the researches in two categories of different variants of MOEA/D are illustrated. Finally, the real world application areas that applied the decomposition approach are mentioned.

Image denoising is an important preprocessing step in two-dimensional gel electrophoresis (2-DGE) that strongly affect spot detection or pixel-based methods. Denoising autoen-coders (DAE) is a new approach in deep learning used in image denoising that has a challenging performance. In this study, DAE technique is applied on 2-DGE images motivated by its ability to learn a robust representation to partially corrupted input. DAE is applied on over than 300 real gels got from LEeB 2-D PAGE database. To validate the efficiency of this technique three indicators are used; Signal-to-noise ratio (SNR), False discovery rate (FDR) and spot efficiency. The average results before denoising are 0.6332 for SNR and 71.05 for spot efficiency. Whereas, the average results after DAE are 61.3317 for SNR, 99.9944 for FDR and 88.4 for spot efficiency. Moreover, DAE outperforms the denoising wavelet by 1.75 %.

The recent development of machine learning and deep learning techniques for medical image analysis has led to the development of intelligent diagnosis systems that can help doctors make a better diagnosis to the patients' diseases. In particular, skin diagnostics is a field where these new techniques can be applied with a high rate of accuracy. This study aims to enhance the accuracy of skin lesions classification based on two factors. The first is deeply trained all layers of implemented pre-trained models. Whereas, the second is balancing the number of images within the seven classes of dataset used. The state-of-the-art convolutional neural networks MobileNet and DenseNet-121 were trained on HAM10000 dataset. The two models pass through three phases; preprocessing, training and evaluation. Firstly, the dataset is down sampled, splitted and augmented to resolve misbalancing problem. Then, both models are deeply trained and finally they are evaluated against baseline models without balancing the classes. Multiple metrics were used to evaluate our models; precision, recall, F1-score, specificity and ROC AUC. In addition, the micro-average and macro-average of all previous metrics to extend to multi-classification. The accuracy of MobileNet and DenseNet-121 reach 82.6% and 71.9% on unseen testing images, respectively on the original dataset (i.e. before balancing the dataset). Whereas, they reach 92.7% and 91.2% on unseen testing images, respectively after balancing the dataset. This enhancement proves the necessity of existence of balanced dataset for training, to have better performance. Furthermore, MobileNet after balancing dataset has out performed the highest accuracy of ISIC 2018 challenge on the same dataset by 4.2%. For that, this model is the recommended one as it is a light-weight model, suitable for mobile applications used by dermatologists and its accuracy is comparably equal to DenseNet121.

Fault localization is considered one of the most challenging activities in the software debugging process. It is vital to guarantee software reliability. Hence, there has been a great demand for automated methods that can pinpoint faults for software developers. Various fault localization techniques that are based on graph mining have been proposed in the literature. These techniques rely on detecting discriminative sub-graphs between failing and passing traces. However, these approaches may not be applicable when the fault does not appear in a discriminative pattern. On the other hand, many approaches focus on selecting potentially faulty program components (statements or predicates) and then ranking these components according to their degree of suspiciousness. One of the difficulties encountered by such approaches is to understand the context of fault occurrence. To address these issues, this paper introduces an approach that helps in analyzing the context of execution traces based on control flow graphs. The proposed approach uses the edge-ranking of basic blocks in software programs using Dstar that proved to be more effective than many fault localization techniques. The proposed method helps in detecting some types of faults that could not be previously detected by many other approaches. Using Siemens benchmark, experiments show the effectiveness of the proposed technique compared to some well-known approaches such as Dstar, Tarantula, SOBER, Cause Transition and Liblit05. The percentage of localized faulty versions versus the percentage of code examined is taken as a measure. For instance, when the percentage of examined code is 30%, the proposed technique can localize nearly 81% of the faulty versions, which outperforms the other four techniques.

Image denoising is the technique of removing noise or distortions from an image. During medical image acquisition, random noise is added, which results in a lower contrast in those images. For that, image denoising is a crucial task for medical imaging analysis. In this study, a denoising system using three heterogeneous medical datasets is proposed based on stacked convolutional autoencoder (SCAE) technique. To validate its efficiency, different evaluation metrics are used, such as mean squared error (MSE), Peak signal-to-noise ratio (PSNR), contrast-to-noise ratio (CNR), structural similarity index measure (SSIM) and cross correlation (CC). The proposed denoising system gives good results among the medical and microscopic datasets that are used for training. The best average results obtained are 0.0039 for MSE, 24.07 for PSNR, 0.1220 for CNR, 0.85 for SSIM, and 0.6358 for CC. Then, the proposed SCAE denoising system was applied to the LECB 2-D PAGE database for denoising real 2-DGE images. The results of denoising 2-DGE images are evaluated by MSE, spot efficiency, false discovery rate (FDR), and signal-to-noise ratio (SNR). The best average results for 2-DGE images are 0.014 for MSE, 75 spot efficiency, 36.3 for FDR and 18.41 for SNR. The proposed system has enhanced the denoising of 2-DGE images by 0.9% to 17.6% when compared to other techniques.

Protein structure prediction is a very challenging problem in drug discovery and computational biology. To achieve better multi-class classification model for fold recognition problem, a combination between semi- supervised and boosting techniques is proposed into Semi- supervised Boosting Fold Recognition (SB-FR) algorithm. In addition, a testing method "TreeTest" is introduced for improving the overall accuracy of SB-FR algorithm. To benchmark the performance of the proposed SB-FR algorithm, a famous challengeable "Ding and Dubchak" dataset is used for training and testing. In addition, different parameters are applied to the same random sets of labeled and unlabeled sequences. To benchmark the "TreeTest" testing method, All-versus-All (AvA) testing method is used for comparison. Finally, using the proposed SB-FR algorithm along with the proposed "TreeTest" method for fold recognition multi-class classification, a 5.6% improvement is recorded in the overall accuracy for three- class and 8.2% for five-class classifications when compared to the base classifier. 

The Hepatitis C virus nonstructural 5A (NS5A) protein is a hydrophilic phosphoprotein with diverse functions. The domain assignment of NS5A had been refined using a systematic in silico bioinformatics approach using DOMAC, the protein is divided into three domains and domain III is subdivided into two subdomains using ProDom and SSEP servers. The fold structure for domains II and III were predicted using the meta-server 3D-Jury. Scanning motif databases (SMART, BLOCKS, and PROSITE) gave new motifs. Two important motifs, the interleukins 1 and 8 interaction motifs, relating to NS5A function in inducing the interleukin 8 promoter, were discovered from the BLOCKS scan. Protein-protein interaction motifs were predicted as hot loops and disordered regions, corresponding to binding regions with the ds-protein kinase R, viral polymerase, and Src homology 3 signaling proteins binding motif. Other hot loops were predicted in the V3 region and in the single-stranded DNA-binding protein motif. The different mechanisms by which the NS5A protein leads to immune system signaling dysfunction points to the natural genetic engineering of this protein.

A methodology for elucidation of structural, functional, and mechanistic knowledge on promiscuous proteins is proposed that constitutes a workflow of integrated bioinformatics analysis. Sequence alignments with closely related homologues can reveal conserved regions which are functionally important. Scanning protein motif databases, along with secondary and surface accessibility predictions integrated with post-translational modification sites (PTMs) prediction reveal functional and protein-binding motifs. Integrating this information about the protein with the GO, SCOP, and CATH annotations of the templates can help to formulate a 3D model with reasonable accuracy even in the case of distant sequence homology. A novel integrative model of the non-structural protein 5A of Hepatitis C virus: a hub promiscuous protein with roles in virus replication and host interactions is proposed. The 3D structure for domain II was predicted based on, the Homo sapiens Replication factor-A protein-1 (RPA1), as a template using consensus meta-servers results. Domain III is an intrinsically unstructured domain with a fold from the retroviral matrix protein, which conducts diverse protein interactions and is involved in viral replication and protein interactions. It also has a single-stranded DNA-binding protein motif (SSDP) signature for pyrimidine binding during viral replication. Two protein-binding motifs with high sequence conservation and disordered regions are proposed; the first corresponds to an Interleukin-8B receptor signature (IL-8R-B), while the second has a lymphotoxin beta receptor (LTβR) high local similarity. A mechanism is proposed to their contribution to NS5A Interferon signaling pathway interception. Lastly, the overlapping between LTβR and SSDP is considered as a sign for NS5A date hubs.