Classic British translation is an important way to "go global" in Chinese culture，Reader comments are the final criteria for translation。The research trend of the current English translation reader's viewpoint is: under the theoretical guidance of readers，Introduction of natural language processing technology and text mining technology，Expansion of the Study of Readers' Studies，Obtain large -scale review corpus through the Internet environment，Automatic mining and quantitative analysis of the comments，Sort out the key topic of the reader's comments，so that the system can deepen valuable viewpoint information，Provides accurate and reliable data analysis for the reader's comments。Comprehensive use of natural language processing technology、Cross -disciplinary research of machine learning methods and semantic analysis，It is an effective way to excavate the British translation of classics，The specific implementation path is as follows。

　　Bet365 app download The main research object with readers from the United States and British readers，Collect review corpus through Amazon Overseas Website API interface，Standard experimental corpus of the British translation book comments in classics，Use natural language processing technology to stop using words and noise data、Word -based labeling and other pre -processing，Analysis of sentences in the context environment、Refers to the dissolution and omittime recovery，Establishing a classic British translation comment corpus。

　　Second，Extract the theme words of the British translation of classics and modifications，Sort out the focus of the reader's comment。Short -length of the text of the online comment、Features such as sparse features，Extracting theme and viewing relationships through technologies such as analysis and semantic analysis of syntax and semantic analysis，Discover the fixed combination mode of the topic and modification language，Analysis and comment on the topic of review themes，The text mode of identifying the corpus of the comment，Automatically extract the explicit theme in the comments，The style of the translation、Structure、Language style, etc.，Inquiry The hidden theme of the review of the commentary corpus but can be obtained through semantic reasoning。

Semantic analysis and domain knowledge representation is the key to improving the accuracy of network review viewpoint。The basis of semantic analysis is vocabulary representation，Use word vectors to represent in the classification task of emotional vocabulary polarization、Feeding neural networks and convolutional neural networks and other deep learning algorithms，Effectively improved the analysis of emotional vocabulary polarity、The accuracy rate of the task of semantic analysis。

Introducing the field of knowledge base analysis The context can understand the emotion that the reader really wants to express，The basic task of improving the knowledge base of the field is the knowledge map supplement，Existing knowledge map supplementation full algorithm consumes a long time、limited accuracy，Interdisciplinary deep learning algorithm is an effective research approach to solve this problem。

　　third，Differential view of emotion polarity，Establishing a classic British translation comments: Abstract。Comments Emotional Poor judgment is the key step of exploring potential views and attitudes。From the perspective of machine learning，Emotional polarity recognition can be regarded as multiple categories、Single label text classification task。Machine learning classification algorithm combined with emotional dictionary，Emotional polarity that can effectively mark viewpoint modification words，Provide objective data for the positive and negative review viewpoints of quantitative research classics and English translation readers; combined with cluster algorithms, we can find the internal connection and objective laws between the themes of the review; Decision，Being able to accurately analyze the theme of the review perspective and emotional polarity; the topic of the theme of the perspective of machine learning and emotional dictionary for formation，Inquiry contains the potential attitude of views in online comments，It can help translators and publishers accurately grasp the reader's positive and negative comment on the reader based on credible data。

　　Fourth，Deep mining comment semantic theme，Get hidden readers' perspective。Classic British translation review viewpoint mining must be based on translation style、Victory such as words such as words，Internal connection and importance of the theme of the point of view from the global situation。Internet comment noise、Expressive ways、Large corpus scale、Disclosure of views。In order to sort out the focus of the reader's attention，The theme model that needs to be built with deep semantic mining，Reveal the theme of the comments at the semantic level，Map the high -dimensional evaluation paper to the low -dimensional theme space，Make it better explanatory，Multi -dimensional analysis and excavation hidden value theme，Knowledge in the field of field，Classified the subjective words extracted，Draw the theme words and conjugate clustering spectrum，Visual similarity mapping technology and weighted module parameterization algorithm presents the theme cluster that overseas readers pay attention to，Combined with the centrality of the knowledge network node to present the key theme of various clusters，It can break through the original book review theme 囿 The limitations of subjective analysis and small sample data，Sentences with a more widely covered from the complicated comments、The hidden knowledge of the theme vocabulary is more abundant。

　　Fifth，Text visualization analysis，System analysis reader comments。Expressive viewpoints in the integrated perspective and hidden perspectives in theme models，According to the equivalent of semantics、Level and Conditions，Merge on the theme words、Description and representation of the upper and lower positions or correlations; sort the theme words according to the importance; summarize which translator the readers are for、What themes of the translation are compared; based on the perspective of theme clustering category, compare the evaluation of readers who are currently widely accepted，Digging of the deep -seated reasons for the best -selling of classics and English translation works; analyzes the distribution of theme words and modified emotional poles，Understand the specific attitude of foreign readers on specific translators or translations，For the translators and publishers to further understand the needs of readers to provide scientific and reliable basis。It can be further adopted by statistics to show the theme words in a theme word cloud.，Sort the summary of the subjective according to importance。Analysis of semantic relationship between theme theme，Semantic description of the explicit comparison relationship in network comments，Calculate the similarity between the topic clusters，Deeply -based language analysis Motor semantic clustering for evaluation views，System analysis reader comments。

　　Sixth，Adapt to multi -language cross -domain environment，Welcome to international challenges。The internationalization characteristics of the Internet determine in multi -language、Research and English translation comments in cross -domain contexts are particularly important，Sentence analysis、Basic analysis methods such as emotional polarity judgment are highly related to the field of language and environment problems，The emotional characteristics of data in different fields are not exactly the same，Emotional prediction model trained in data in a certain field，Usually not directly used in other fields。With the number of users' comments and the number of various fields continued to increase，It takes a lot of time and resources for individual training models for all fields。

Cross -domain emotional classification through the knowledge improvement target field through the relevant source field，Specific realization of migration learning through similar fields or field adaptation models，For example, the emotional classifiers obtained by comments in the field of book reviews，Migration bet365 best casino games or adaptation to the field of digital video CD，Label time and resource for saving the comments in this field。Comment emotion usually has a problem of dispersion in different fields，Expressive emotions such as "Readability" and "Thinking" in the field of books，Use "flat" and "no plot" to represent negative emotions; and in the field of digital video CD，Usually "high definition" and "smooth" and other positive emotions，Use "fuzzy" and "scratches" to represent negative emotions。Due to differences between fields，Emotional classification model trained in the source field，It often performs poorly when it is directly applied to the target field。Using deep learning method，It can provide solutions for emotional characteristics drift in cross -domain environment，The difficulty that needs to be solved is how to handle the text that is rich in semantic short text。

Cross -language emotional analysis is an analysis of emotional tendencies of target language text using source language text，Specific implementation can be based on resource migration and joint learning -based method。Resource migration method is difficult to implement due to the different language and corpus injection system，Based on the joint learning method, it mainly depends on machine translation，The quality of the translation results is greaterly affected。In recent years，Deep learning has become a hot spot for cross -language emotional analysis research，At present, it mainly focuses on the thick particle size level，Cross -language fine -grained emotional analysis needs further research。

(The author is the head of the National Social Science Fund project "Dian British translation of foreign readers online review viewpoint mining research" in charge、Professor of Dalian University of Foreign Languages)