基于用户行为的教学视频内容质量评价方法

针对目前网络教学视频内容质量评价以主观方法为主,缺乏客观的质量界定标准的问题,提出一种基于用户观看行为的网络教学视频质量评价方法.该方法首先采集单个用户观看某个网络教学视频的行为数据,并对数据进行标准化处理;然后根据视频质量评价标准,实现数据标签化;再通过全连接神经网络,利用Softmax划分单个用户对网络教学视频内容质量的分类;最后,将所有用户观看该视频的分类加权平均后得到对该视频的综合评价.测试结果表明,该模型评价教学视频的准确率为79.5%,分类效果明显,具有较高的实用价值.

Quality evaluation of teaching video content based on user behavior

Currently, the problem that subjective method is the main method for evaluating the quality of online teaching video content, lacking of objective criterion for defining its quality. In this paper, aonline teaching video quality evaluation method is proposed based on user behavior. In this method, the behavior data of single user watching a networked teaching video is collected first and normalized. Data tagging is then implemented based on evaluation standard of video quality and the quality classification of the networked teaching video content is conducted by the single user by means of fully connected neural network and Softmax. Finally, a classification weighted average of the ratings of all users watching this video gives a comprehensive evaluation of the video. The test result shows that the accuracy of this model will be 79.5% and the classification effect will be obvious, having higher practical value.