Financial public opinion perception method based on weighted LDA (latent Dirichlet allocation) topic model

Abstract

The invention discloses a financial public opinion perception method based on a weighted LDA (latent Dirichlet allocation) topic model and belongs to the technical field of natural language understanding and processing as well as network public opinion. Everyday financial public opinions are perceived on the basis of microblog data related to everyday finance and are quantified according to 'everyday financial public opinion comprehensive index'. The 'everyday financial public opinion comprehensive index' is a weighted average of all financial related blog emotion values on the day, and the blog emotion values are a result of text emotion classification of blog content. An SVM (support vector machine) classification model based on weighted LDA is adopted for text emotion classification and adopts the weighted LDA for establishing text represented hidden topic space, objective data indirectly embodying investor sentiment and subjective data directly embodying investor sentiment are organically combined with a new lexical item weight calculation method, and accurate understanding of texts from the semantic level is promoted greatly, so that the text emotion classification effect is better.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

In summary, these methods aim to enhance understanding how people see things like stock prices or social media interactions during their online conversations. By combining relevant factors such as comment count and user feedback about what they have posted onto them over time, we found out who took certain actions earlier were influenced negatively towards those later appearing. These techniques help us better identify patterns related to economic activity and make informed decisions for future transactions.

Problems solved by technology

Technological Problem: The technical problem addressed in this patented text relates to improving the accuracy of predicting economic outcomes through internet media (Blogger) activity collected during an initial phase called web browsing. Current methods involve analyzing irrelevant details like comment marks or mentions made by users who want to see what's happening around them rather than directly measurable factors like income levels. Additionally, there may exist other sources where people act differently towards certain things over time due to belief bias caused by negative feedback provided by previous experiences with similar products or services.

Images