Header Journal2

Albaha Univeresity Journual of Basic and Applied Sciences

Asset Publisher

  • Mohammed Y. Al-Ghamdi
  • Reseived: 8 April 2019 Received in Revised Form: 14 October 2019 Accepted: 23 October 2019
  • doi

In the fast-changing environment led by the rise of the internet and new technologies, e-learning has emerged as the most promising solution for the universities. Perception and attitude of students towards IT application and e-learning technologies are important factors towards the successful development of academic programs. This paper is focused on the adaptation of e-learning as an alternative to traditional learning. Moreover, it looks at the perception of lecturers towards e-learning. The paper investigates two main points: (i) the impact of e-learning on academic performance, (ii) the motivation of the students. Through an exhaustive review of the preceding researches in this domain, it was concluded that e-learning offers several benefits over traditional methods of learning. Lecturers too perceive e-learning in a positive manner as it assists them in a better delivery of lesson contents. Amongst several other benefits, e-learning offers more flexibility and greater channels of communication as compared to traditional classroom-based learning. Students undertaking e-learning have greater internal motivation compared to those opting for classroom-based education. Most importantly, e-learning is shown to have a positive impact on the performance of the student.

  • Mohammad E. Alzahrani
  • Reseived: 26 August 2019 Received in Revised Form: 22 October 2019 Accepted: 18 November 2019

This paper enhances the Convolutional Neural Network (CNN) to increase accuracy in recognizing faces in a crowd. The enhanced CNN is built upon the well-known VGG-19 which a very deep learning CNN technique. Prior to the recognition phase, the image data was preprocessed using standard image processing mechanism. Experiments on existing publicly available Labeled Faces in the Wild (LFW) dataset were conducted. The proposed enhanced-CNN outperforms the other existing methods and achieved up to 99.91% accuracy. The best accuracy was obtained from the experiment on LFW-100 sub-dataset, which has 5 persons with 100 images/person. In future, the proposed enhanced-CNN can be used for the development of Hajj safety alert system.

  • Ayman H. Mansee Doaa M. Abd El-Gwad Abd El‐Salam M. Marei
  • Reseived: 9 April 2019 Received in Revised Form: 15 November 2019 Accepted: 24 November 2019

Different bacterial isolates from two types of soils differentially in atrazine applications history were isolated. Two consortiums from each soil sample (Sandy loam and Clay loam) were isolated using non-selective and then selective media and tested for their atrazine degradation efficacy. Effects of different biostimulators on consortiums growth and atrazine degradation efficacy were also studied. The consortium AyDds isolated from the soil of a long history of atrazine applications by selective media was capable of degrading 82.97% of applied atrazine in 60 days. The growth of consortiums reached the highest value when all nitrogen sources were removed from the media. The rate of atrazine degradation was increased by cells grown in a media supplemented with ammonium instead of atrazine as a sole source for nitrogen.

  • Adil Fahad
  • Reseived: 16 September 2019 Received in Revised Form: 19 November 2019 Accepted: 26 November 2019

A few classification frameworks have focused on a new challenge in data streams, namely, concept-evolution, which occurs when a totally new class/concept appears in the stream. However, the recent class detection techniques dealing with this challenge focus on a single data stream and numeric data, and they have high false detection rates in many scenarios. This paper proposes a new classification technique, called FUzzy sub-clusters for Novel class Detection in multi data streams (FUND), that maintains a set of representatives of each class label, a set of clusters features, as well as extending the boundaries of the sub-clusters to allow the fuzziness nature of the data streams. In particular, FUND applies two similarity metrics, namely density and distance, to overcome the uncertainty issue in data streams and it distinguishes the instances that belong to an existing class from the potential novel class instances. The improvement in the detection relates to the classification accuracy by making sure that an instance of a novel class does not end up being classified as an existing class instance and vice versa. Experimental results show that FUND outperforms state-of-the-art data stream classification techniques in both synthetic and challenging real datasets. The average accuracy improvement is around 30% for FUND-C compared to MineClass and DNTC techniques.

  • Faleh Z. Alqahtany
  • Reseived: 13 October 2019 Received in Revised Form: 3 December 2019 Accepted: 10 December 2019

The quality of water is one of the most essential solicitudes. Heavy metals are main constituents of wastes which have been involved in several metal diseases and food poisoning in man. This study evaluated concentration of iron, lead, cobalt, nickel, chromium, cadmium, vanadium, aluminum, cobalt, selenium, molybdenum, mercury, lithium, manganese, boron, beryllium, zinc, and arsenic in water samples from different locations in Asser province, Saudi Arabia. Results revealed that about eighteen heavy metals were detected in water samples. Two heavy elements, Mercury and manganese have highest concentrations (mg/L) of 0.0086, 0.0034, 0.0021, and 0.1433 in samples from Alshaboa-quarter, Al rakla-quarter, Tathleeth-industry, Tathleeth-Bisha gate and Abha-Alryiad Road, respectively. pH is measured for all samples at room temperature. Total hardness (TH) was determined by a titration method. The detection of heavy metals in samples rendering for consciousness. This is important to dishearten researches which could lead to high metals concentration and mineral toxicity.