The trend of distance learning education has increased year by year because of the rapid advancement of information and communication technologies. Distance learning system can be regarded as one of ubiquitous computi...The trend of distance learning education has increased year by year because of the rapid advancement of information and communication technologies. Distance learning system can be regarded as one of ubiquitous computing applications since the learners can study anywhere even in mobile environments. However, the instructor cannot know if the learners comprehend the lecture or not since each learner is physically isolated. Therefore, a framework which detects the learners’ concentration condition is required. If a distance learning system obtains the information that many learners are not concentrated on the class due to the incomprehensible lecture style, the instructor can perceive it through the system and change the presentation strategy. This is a context-aware technology which is widely used for ubiquitous computing services. In this paper, an efficient distance learning system, which accurately detects learners’ concentration condition during a class, is proposed. The proposed system uses multiple biological information which are learners’ eye movement metrics, i.e. fixation counts, fixation rate, fixation duration and average saccade length obtained by an eye tracking system. The learners’ concentration condition is classified by using machine learning techniques. The proposed system has performed the detection accuracy of 90.7% when Multilayer Perceptron is used as a classifier. In addition, the effectiveness of the proposed eye metrics has been confirmed. Furthermore, it has been clarified that the fixation duration is the most important eye metric among the four metrics based on the investigation of evaluation experiment.展开更多
The underflow concentration prediction of deep-cone thickener is a difficult problem in paste filling. The existing prediction model only determines the influence of some parameters on the underflow concentration, but...The underflow concentration prediction of deep-cone thickener is a difficult problem in paste filling. The existing prediction model only determines the influence of some parameters on the underflow concentration, but lacks a prediction model that comprehensively considers the thickening process and various factors. This paper proposed a model which analyzed the variation of the underflow concentration from a number of influencing factors in the concentrating process. It can accurately predict the underflow concentration. After preprocessing and feature selection of the history data set of the deep-cone thickener, this model uses the eXtreme gradient boosting(XGBOOST) in machine learning to deal with the relationship between the influencing factors and the underflow concentration, so as to achieve a more comprehensive prediction of the underflow concentration of the deep-cone thickener. The experimental results show that the underflow concentration prediction model based on XGBOOST shows a mean absolute error(MAE) of 0.31% and a running time of 1.6 s on the test set constructed in this paper, which fully meet the demand. By comparing the following three classical algorithms: back propagation(BP) neural network, support vector regression(SVR) and linear regression, we further verified the superiority of XGBOOST under the conditions of this study.展开更多
文摘The trend of distance learning education has increased year by year because of the rapid advancement of information and communication technologies. Distance learning system can be regarded as one of ubiquitous computing applications since the learners can study anywhere even in mobile environments. However, the instructor cannot know if the learners comprehend the lecture or not since each learner is physically isolated. Therefore, a framework which detects the learners’ concentration condition is required. If a distance learning system obtains the information that many learners are not concentrated on the class due to the incomprehensible lecture style, the instructor can perceive it through the system and change the presentation strategy. This is a context-aware technology which is widely used for ubiquitous computing services. In this paper, an efficient distance learning system, which accurately detects learners’ concentration condition during a class, is proposed. The proposed system uses multiple biological information which are learners’ eye movement metrics, i.e. fixation counts, fixation rate, fixation duration and average saccade length obtained by an eye tracking system. The learners’ concentration condition is classified by using machine learning techniques. The proposed system has performed the detection accuracy of 90.7% when Multilayer Perceptron is used as a classifier. In addition, the effectiveness of the proposed eye metrics has been confirmed. Furthermore, it has been clarified that the fixation duration is the most important eye metric among the four metrics based on the investigation of evaluation experiment.
基金supported by the National Key Research and Development Program of China(2016YFB0700500)the National Science Foundation of China(61572075,61702036)+1 种基金Fundamental Research Funds for the Central Universities(FRF-TP-17-012A1)China Postdoctoral Science Foundation(2017M620619)。
文摘The underflow concentration prediction of deep-cone thickener is a difficult problem in paste filling. The existing prediction model only determines the influence of some parameters on the underflow concentration, but lacks a prediction model that comprehensively considers the thickening process and various factors. This paper proposed a model which analyzed the variation of the underflow concentration from a number of influencing factors in the concentrating process. It can accurately predict the underflow concentration. After preprocessing and feature selection of the history data set of the deep-cone thickener, this model uses the eXtreme gradient boosting(XGBOOST) in machine learning to deal with the relationship between the influencing factors and the underflow concentration, so as to achieve a more comprehensive prediction of the underflow concentration of the deep-cone thickener. The experimental results show that the underflow concentration prediction model based on XGBOOST shows a mean absolute error(MAE) of 0.31% and a running time of 1.6 s on the test set constructed in this paper, which fully meet the demand. By comparing the following three classical algorithms: back propagation(BP) neural network, support vector regression(SVR) and linear regression, we further verified the superiority of XGBOOST under the conditions of this study.
