The development of intelligent technologies gains popularity in the education field. The rapid growth of educational data indicates traditional processing methods may have limitations and distortion. Therefore, reconstructing the research technology of data mining in the education field has become increasingly prominent. In order to avoid unreasonable evaluation results and monitor the students' future performance in advance, this paper comprehensively uses the relevant theories of clustering, discrimination and convolution neural network to analyze and predict students' academic performance. Firstly, this paper proposes that the clustering-number determination is optimized by using a statistic which has never been used in the algorithm of K-means. Then, the clustering effect of K-means algorithm is tested by discriminant analysis. The convolutional neural network is introduced for training and testing data that are labeled with categories. The generated model can be used to predict prospective performance. Finally, in order to validate the prediction results, the effectiveness of the generated model is evaluated by using two metrics in two cross-validation methods. The experimental result demonstrates that the statistic not only solves the difficulty to determine the clustering number in K-means algorithm from an objective and quantitative point of view, but also improves the reliability of prediction results.

The data in modern educational information systems are not given enough attention and are not fully utilized. Therefore, the motivations of our study are to preliminarily explore learning behavior patterns by applying process mining to educational datasets, and construct prediction models based on previous learning behavior. The data in modern educational information systems can be used by teaching managers to analyze various aspects of the educational process from different perspectives. We prepare to choose three datasets randomly which include student number, courses and grades attributes from a university's educational information systems. This paper firstly applies system clustering to give an overview of students' academic performance, and roughly determines clustering number. In order to ensure the accuracy which is relevant to the analysis of students' learning behavior patterns, a semi-biased statistic is proposed to quantitatively determine clustering number. Then, the data are clustered by fast clustering algorithm, and the clustering effect is cross-validated which is aimed at accurately analyzing the behavior patterns of student groups and using data visualization technology to visualize different student groups. Finally, the support vector machine is used to construct a classifier for predicting the learning behavior pattern, and the parameters in the support vector machine are optimized by Bayesian optimization, genetic algorithm optimization and whale optimization respectively. The research found that: 1) In the equal test of the group mean, when the significance of most courses is less than 0.05, it means that there is a significant difference among different categories. In this case, using the semi-biased statistic proposed in the paper is helpful to improve clustering effect. 2) The better the students learn, the better the clustering effect of the category which they belong to is. 3) Whale optimization algorithm works best.INDEX TERMS Educational process mining, learning behavior patterns, student profiles, support vector machines, system clustering, the semi-biased statistic.

No abstract

Currently, as a universal clean energy, natural gas plays a greater role in industrial and civil energy consumption than it has previously. Any insufficient supply scenario has a severe impact due to the increasing use of power plants, chemical engineering, industrial production, and public sectors. It is essential to develop a methodology for analyzing gas supply insufficiencies that are caused by pipeline network malfunctions. This paper introduces a systematic method for evaluating the natural gas supply reliability based on the pipeline network. Primarily, the reliability of each unit in the pipeline network is derived from multi-variant distribution principles to initiate topological structure analysis carried out in the real pipeline network. Afterwards, the Monte Carlo simulation shows the random status of the topological network based on preconcerted failure distributions of facilities and pipes rather than estimating the reliability directly. Because the current transmission capacity is possibly excessive relative to the transmission task, both designed capacity and current supply capacity require stochastic simulations. After stochastic simulations of the market demand, a feasible random transmission requirement and a certain structure of the topological network are obtained from random simulations to calculate the total transmission capacity. Ultimately, according to the supply insufficiency level, there are deployable measures that could eliminate this influence.