Research And Implementation Of Text Orientation Comparison Method For Social Network

With the popularity of mobile networks,social networks represented by Weibo have gradually become an important channel for people to obtain information and express opinions.Texts in social networks are usually short,with large amounts of data,and has important value of data mining.It is an important data source for public opinion analysis.The comparison of text orientation is to determine the orientation of the text to be analyzed by comparing the texts with known orientations,which is a means of text orientation analysis.Text orientation analysis is a research hotspot in the fields of natural language processing and public opinion analysis.Most of the previous studies have neglected the potential emotional factors contained in the context of text,or use neural network,deep learning and other mechanisms,which makes its principle complex and unsuitable for processing short texts such as Weibo.Aiming at these problems,this paper proposes a text orientation comparison method,and verifies its effectiveness through experiments.Finally,the system design and implementation are carried out in combination with Hadoop distributed computing platform.The main work of this paper includes:1.Aiming at the short length and large amounts of data from social network such as Weibo,this paper proposes a text orientation comparison method,which is to compare the semantic similarity between the texts to be analyzed and the standard texts representing positive,negative and neutral,and then divide them into a more similar category.In addition,the principle of selection of standard texts has also been proposed.2.1n the aspect of semantic similarity calculation,the concept of fusion semantic similarity and its calculation method are proposed,which is to calculate the text similarity in the relevant subspace and the text sentiment similarity in the emotional subspace,and combine the two results.This method not only utilizes the explicit emotional factors in the emotional and opinion words,but also utilizes the text context with implicit emotional factors,which makes this semantic similarity more suitable for text orientation analysis.3.Combined with the Hadoop distributed computing platform,the text orientation comparison method using fusion semantic similarity is implemented systematically.Data is acquired by the data collection layer through the Weibo API and stored in a local database.After the texts art preprocessed in parallel,combined with the orientation comparison method proposed in this paper,their orientations are analyzed in parallel,which can increase the effectiveness of data processing.The result display layer retrieves the text orientation analysis results and displays them.