Toward generating a new video education lectures dataset and performance comparison with various machine learning algorithms

Abstract

During the last decade, Data has a significant role in analytic processes, applications, and numerous areas of the human activities. With the growth in data complexity and size, we come across a vital necessity for tools promoting the explanations and understanding over data-associated procedures. Due to the lack of adequate dataset, anomaly-based approaches in education resources datasets which are monopoly to the universities or educational institutions. The objective of this research is to build a new dataset called video educational lectures dataset due to the lake of educational resources datasets this dataset designed for postgraduate students, then test the performance of (Video Lectures) dataset by applying a set of algorithms of machine learning for the tasks of data mining, the different classification methods, are: J48, Decision Tree, OneR, Bayes Network, Random Forest, LMT, Random Tree, LWL, REP, and Navies Bayes algorithm, after that the new dataset will be tested and evaluated by compare it with a public dataset (Good Book) dataset through using the same classification algorithms. The several measures which are considered for the analytical purposes are Sensitivity, Accuracy, Precision, F-measure, and False positive Rate. All the results of those parameters are taken for all of the described methods of classification. Finally, results have been provided in tabular way to test and evaluate the performance of the Video Lectures dataset, and other tabular form to facilitate comparative between the Video Lectures and Good Book datasets that proved the Video Lectures dataset get better Accuracy percentage than Good Book dataset.