Research On Blog Post Quality Evaluation Based On Amr And Elegant Sentence Recognition

With the rapid development and popularization of the Internet,various social media have emerged one after another.Social platforms such as Weibo,We Chat Moments,QQ News,etc.are favored by users for their convenient publishing,real-time information,concise content,and easy user communication.The number of its users has exploded in recent years,and the amount of data generated has increased at an exponential rate.However,with the rapid development of social platforms,the phenomenon that netizens only demand quantity but not quality has led to serious information quality problems.Coupled with factors such as the imperfect monitoring function of the platform and the fragmentation of information,the development of the platform is very important for the development of the platform.Maintenance work has caused trouble.Among them,the quality of text information is the focus of attention.Although massive data resources support the development and research of text technology,its value density is too low,the data contains a lot of repetition,noise and junk data,and the quality of text is even more uneven.Evaluating the quality of short texts is a key issue for many applications(such as recommendation systems and online searches to find high-quality articles and filter out low-quality articles).It has become an urgent problem to be solved that how to rapidly determine the quality of the text information.This article conducts quality evaluation research on short texts of microblogs.Foreign research techniques on text quality have become mature and have reached a practical stage.However,the domestic research on text quality evaluation started late,and Chinese is different from English,so Chinese can't just Simply analyze its literal meaning,such as evaluating the quality of a sentence based on shallow features such as single word,single word,sentence length,number of words,etc.,and deeply analyze the sentence based on its grammar,semantics,and pragmatic functions.quality.This article mainly studies text quality from the perspective of graceful sentence recognition.At present,there are few researches on graceful sentence recognition,and it is difficult to extract its features.In recent years,BERT and its variant models have achieved good results in text classification tasks.Existing experiments have shown that the BERT model can extract the grammatical and semantic information of sentences.Through in-depth research on the BERT model,it is found that it is highly sensitive to the recognition of beautiful sentences.Through a large amount of labeled corpus for pre-training,the BERT model can perform in-depth analysis of the structure of the beautiful sentence,and capture its grammar,Semantic information and structural features.Through in-depth analysis of the experimental results of the AMR quality evaluation model,it is found that it has certain defects.Finally,a method combining AMR quality assessment and elegant sentence recognition is proposed.First,the AMR quality evaluation model is used to evaluate the completeness of sentence structure and the closeness of sentence sequence,and conduct an in-depth analysis of the short text of Weibo from the grammatical level of syntactic structure and modified semantics.Secondly,use a large number of beautiful sentence corpus to train the BERT model,and then use the trained BERT model to evaluate the sentences with lower scores on the AMR quality evaluation model.Finally,according to the comprehensive evaluation of the AMR quality evaluation and the beautiful sentence recognition model,the quality of the blog post is scored.There are three levels: high,medium,and low.