The Construction of Teaching Mode Based on the Teaching Data Classification by the Baidu API and Deep Learning Models

This paper introduces the construction of teaching mode that applies the deep learning technology to the teaching data classification. Our method uses Baidu API's figure and OCR recognition service that are to output the general information of the teaching data. Then, we use deep learning models to classify the specific information. We fuse these to correctly classify the teaching information. We filter the new labels based on the objective of teaching mode and then use the corresponding samples to train the new models, which is to achieve the scalability. We utilized some public datasets for the experiment, and then we apply our method to the collection of teaching data to prove the scalability.