Research On The Construction Of Metacognitive Ability Model For Online Learning

Prosperity of education will make the country prosperous and strong.With the in-depth development of the "Internet + Education" model,online learning has gradually become the mainstream form of current education.Online learning breaks the boundaries of time and space,provides learners with high-quality learning resources and free learning space,and enables learners to realize their desire to learn anytime,anywhere.However,it also puts forward higher requirements for learners’ autonomous learning ability.Learners need to have a clearer understanding of the individual and the learning environment,be able to grasp their own learning state and cognitive level,and rely on metacognitive ability to understand cognition.The process is reasonably planned,monitored and adjusted to achieve the goal of online learning.Online learning behavior contains a wealth of learning thinking processes and learning states.Deep mining of learning behavior data helps to analyze learners’ cognitive level and metacognitive abilities in real time,helps learners understand themselves clearly,and is beneficial to teachers implement hierarchical teaching for learners at different levels to promote effective online learning for learners.Real-time analysis of the learner’s cognition level helps to monitor the learner’s cognitive state and help the learner have a clear position on himself.Traditional cognitive level analysis methods mainly rely on manual analysis or traditional machine learning techniques.These methods are time-consuming and labor-intensive,and require a lot of feature engineering.Unstructured text data in the forum area,as an explicit form of learners’ thinking expression and knowledge processing,contains rich semantic information,which can often reflect hidden cognitive levels.Based on Bloom’s cognitive target classification theory as the analysis framework,this paper proposes an online learner’s cognitive level evaluation model based on deep neural networks to realize the preliminary exploration of real-time analysis of cognitive level by deep learning technology.Based on the proposed method,the interactive text data released by 9167 learners in the online learning platform is used as an example to analyze.The results show that the constructed deep neural network method can effectively applied to the automatic evaluation of the cognitive level of online learners.Metacognitive ability belongs to the internal high-level thinking process,which is not easy to observe and obtain,and it is usually difficult to judge directly based on the learner’s explicit situation.Existing studies usually use questionnaires,vocal thinking,and self-report methods to quantify metacognitive ability.Although these methods can measure metacognition more accurately to a certain extent,it is difficult to measure metacognition in real time based on actual conditions.Unsupervised online learning behavior is basically carried out under the guidance of its own metacognition.Metacognition will affect the type of learning behavior and the order in which it occurs,resulting in different behavioral data.This paper adopts deep cyclic neural network sequence modeling,in-depth mining and analysis of behavioral data such as learner behavior sequence and interactive discussion,constructs learner metacognitive ability model,and conducts experimental analysis with 151 online learners as experimental subjects,the accuracy rate of the experimental results reached 85.20%,indicating that the model can more accurately quantify learners’ metacognitive ability,verifying the effectiveness of the method in this article,and expect to be online The measurement of learners’ metacognitive ability in the learning process provides technical reference,so as to cultivate and improve the metacognitive level of learners,and promote more effective and personalized online learning.