Article Preview
TopIntroduction
The inclusion of mobile terminals technology into face-to-face teaching boosts the modes of blended learning which thoughtfully integrates the face-to-face learning with online experiences (Bela et al., 2014). In particular, with convenient mobile terminal and online platforms provided to support learning activities, the blended learning modes with technological challenges attracted extensive attention in educational institutions and research communities. How to incorporate the advantages of online learning with offline teaching (Aparicio et al., 2021) in order to improve learning experiences and to promote teaching quality has become an open issue and continued to be an active research topic.
The modes of blended learning are confronted with serious challenges in practical vocational education (Rasheed et al., 2020). For example, some teachers are disturbed by the ineffective connection between online learning and offline teaching activities, which results in insufficient teacher-student communication and student-student cooperation, and lead to unsatisfactory teaching effect from an educator’s perspective and gains depressing learning outcomes from a learner’s perspective (Jia et al., 2016). In order to give full play to the advantages of blended learning in vocational education, the educators and teachers present several solutions from different aspects (Shi et al., 2022; Yu et al., 2020; Zheng et al., 2021a). Zheng et al. applied emotion recognition based on electroencephalogram (EEG) signals to identify the effect of human-computer interaction in online learning (Zheng et al., 2021b). Chen et al. introduced the eye-tracking technology in course recommendation and encouraged the learners to participate in course interaction (Chen et al., 2020). Yu et al. proposed multi-granularity-based Bi-LSTM model with an attention mechanism in Chinese Q&A systems to analyze the contexts in teacher-student interaction (Yu et al., 2020). Liu et al. utilized emotional and cognitive engagement to predict the learner’s achievement based on the learning behaviors in the MOOC forum discussions (Liu et al., 2020). Obviously, even though complex technology and costly devices are available, limited and unsatisfactory effect is achieved dues to the technological challenges encountered in blended learning, such as the inconvenience of learning tools on portability, the inflexibility at learning evaluation, the complexity of learning resources, the boring in online learning and so on (Lin et al., 2016; Yang et al., 2021).
To address the difficulties in present blended learning, the paper has investigated diverse learning systems with different platforms and various technologies to improve teacher-student interaction and promote learning outcomes (Yin et al., 2021). On one hand, with big data available and multi-dimension features extracted, the deep learning models are widely used in various fields benefited from their overwhelming performance (Yu et al., 2022a; Yuan et al., 2021). However, as the samples from vocational education school is limited and the data scale is small, it is proved that the generally used deep learning models are slightly inferior to traditional machine learning methods in performance (Yu et al., 2021; Zheng et al., 2021c). Since a deep learning model contains more parameters in model training, it would spend much more time in execution than a traditional machine learning one does. Moreover, the interpretability of deep learning models hampers their prevailing in practical applications, especially in the field of vocational education. Therefore, the traditional machine learning methods are the dominant ones adopted in the proposed learning system in the paper which reveal remarkable performance with a relatively small amount of data available.