Research And Implementation Of Adaptive Learning System Based On Equipment Maintenance

The traditional virtual maintenance training system treats all students equally, ignoring the individual differences of the trainees. In the virtual maintenance training system, the use of adaptive learning can not only cross the space distance, but also to achieve "individualized" to improve the learning efficiency and learning quality of students. The adaptive learning system can be regarded as a kind of recommendation system,which completes the mapping of "knowledge" to "student". The research content involves student model, domain knowledge model,recommendation strategy and adaptive test. The details as follows:In this paper, we extract term of Mechanical Engineering in random forest model and extract relation between terms in machine learning algorithm. Taking textbooks of Mechanical Engineering as input, the paper extracts the TF-IDF, location, part of speech and so on, and then uses the random forest classification model of word, two words and three words to complete the extraction of mechanical engineering terminology. The relation extraction algorithm is consisted of the algorithm based on topological sorting to extract "Predecessor-successor" relationship and Apriori algorithm to extract Association relationship, as well as an algorithm based on rule to extract Classifying relationship. Experiments show that the accuracy rate, the recall rate and the F1 value of the stochastic forest term extraction algorithm respectively reach 0.51, 0.53, 0.52, it is better than the single TF-IDF, and the decision tree; the accuracy rate, recall rate and FI value of the "precursor-successor" relation of the relation extraction algorithm are extracted to 0.88, 0.93 and 0.904 respectively, and it can extract a certain number of association relations.In this paper, a content navigation algorithm based on knowledge graph and machine learning is proposed. In the "cold-start",this paper achieves content navigation by expanding the border nodes, correct the error node and review the pioneer node in the way of contrasting knowledge models. With the increase of data, recommendation algorithms based on demographics and collaborative filtering are used to implement content navigation using student models and student learning records.After six students' study feedback, the content navigation algorithm in this article can improve students' learning efficiency and quality.This paper realizes the adaptive test algorithm based on cognitive diagnosis model. This algorithm is based on the knowledge level of the trainees, the titles' attributes and so on, so as to realize the adaptive test of the learners. Experiments show that the cognitive diagnosis model used in this paper can quickly find the weak points of students and realize the diagnosis of knowledge structure.