Construction Of Sentiment Index Based On Internet Lending News And Applied Research

As a product of the early exploration of China’s financial technology,Internet lending industry has experienced a process from prosperity to decline.Although the industry has been cleaned-up,its development process helps to deeply understand the development law of financial technology and instruct to develop in the future.Financial market sentiment research is an important field of behavioral finance.Driven by big data,researchers began to extract emotional information from media news,company reports and social networking platforms.For the informal financial industry such as P2P,there is no endorsement of financial license.Even if the platform has excellent qualification,it is easy to be affected by the overall industry sentiment or public opinion environment.Therefore,it is of positive significance to capture the industry sentiment for the operation of P2P platform.Firstly,we construct a sentiment dictionary for Internet finance,which expands the existing sentiment dictionary and provides a better classification basis for sentiment analysis.Based on the sentiment dictionary and machine learning method,this paper improves the sentiment classification method at the sentence level.When using the word vector model to construct the features of sentences,it gives different weights to the sentiment information before and after the transition words,and highlights the weight of sentiment words.This method can improve the accuracy of sentiment classification of long news text and complex sentence structure text.This paper uses the innovative sentiment analysis method to construct the online loan news sentiment index,and further refines it into local level sentiment index and platform level negative news identification variables.Granger test results show that the online loan news sentiment index of P2P industry and industry failure rate are significant.Furthermore,the survival analysis model was used to explore the influence of sentiment variables on the survival risk rate of the platform.The results show that the effect of online loan news sentiment index is significant.When the sentiment variable rises by one unit,the death risk rate of the platform increases by 3%,and the local news sentiment variable has a greater and more direct impact on the platform.In addition,the negative news of the platform has a direct and fatal impact on the platform,and the risk rate of the platform with negative news suddenly becomes 3.49 times of the platform without negative news.In addition,there is a significant contagion effect among industries.When the number of mine blasting platforms in the same province increases,the survival risk of platforms will increase significantly.Finally,this paper constructs the risk early warning model of online lending platform,and finds that the model can more effectively predict the closure time of the platform after introducing the industry sentiment variable.