基于深度学习的影视剧评论情感分析与应用

基于深度学习的影视剧评论情感分析与应用摘要国家与国家之间的交流包括许多方面，分别是：经济、文化等。如果一个国家的文化在世界上具有较强的影响力，对该国文化自信等方面的树立有非常积极的影响，因此，我们国家一直在积极对外传播自身的文化从而提升本国的文化软实力。而影视剧就是传播文化文化中非常重要的途径中的一种。中华五千年的文化在我们国产的影视剧中有所反应，因此可以将其看作我们国家文化在世界上的延申，所以国产剧对我们国家的文化输出而言极为重要。随着互联网技术的发展，用户在各大平台上观看影视剧的时候可以通过平台给出自己对影视据的评价。在观众给出的评价中，包含着观众对观看影视据的真实感受、观点态度等，影视剧创作者能够通过对用户评论中包含的情感进行分析，并根据用户的评论中反映出的情感倾向作为影视剧创作方向把握的一项重要因素，对国家影视剧未来的发展有积极的影响，促进国内的影视剧对外有更广阔的输出空间。因此，本次研究根据深度学习模型，对评论中的句子进行分析，利用模型强大的信息判别能力及特征自学能力，对句子噪声的容忍度进行分析，对影视剧观众评论中的情感分析有一个更深的了解。本次研究利用深度学习模型，并利用CNN技术提取出评论句子中的目标结构特征及BiLSTM技术从远距离语句中提取出其中语义间相互的依赖关系，从而加你BiLSTM-CNN混合模型，从而对影视剧评论中的情感有一个深入的认识，为影视剧未来的发展提供一定科学数据支撑，进而推动国家影视剧的发展，提升国家对外输出的软实力。关键词：深度学习，影视剧，评论，情感分析1Emotional analysis and application of film and television dramacriticism based on deep learningABSTRACTThe exchanges between countries include many aspects,including economy,culture and so on.If acountry's culture has a strong influence in the world,it will have a very positive impact on theestablishment of its cultural self-confidence.Therefore,our country has been actively spreading its ownculture to the outside world,so as to enhance its cultural soft power.Film and television drama is one of themost important ways to spread culture.China's 5000 year old culture has reacted to our domestic film andtelevision dramas,so we can regard it as the extension of our national culture in the world,so domesticdramas are extremely important for our country's cultural output.With the development of Intemettechnology,users can give their own evaluation of film and television through the platform when watchingfilm and television dramas on major platforms.The evaluation given by the audience includes theaudience's real feelings,views and attitudes towards the basis of watching films and television.The creatorsof film and television dramas can analyze the emotions contained in the user's comments and take theemotional tendency reflected in the user's comments as an important factor to grasp the direction of filmand television drama creation,which will have a positive impact on the future development of national filmand television dramas and promote a broader export space for domestic film and television dramas.Therefore,according to the deep leaming model,this study analyzes the sentences in the comments,analyzes the tolerance of sentence noise by using the powerful information discrimination ability andfeature self-study ability of the model,and has a deeper understanding of the emotional analysis in thecomments of film and television viewers.This study uses the deep leaming model,CNN technology to extract the target structural features inthe comment sentences,and bilstm technology to extract the mutual dependencies between semantics fromthe long-distance sentences,so as to add you bilstm CNN hybrid model,so as to have an in-depthunderstanding of the emotions in the film and television drama comments,provide some scientific datasupport for the future development of film and television drama,and then promote the de