Donation-Based Crowdfunding Title Classification Based on BERT+CNN

Abstract

With the rapid development of the Internet, more and more donation-based crowdfunding information is published and forwarded on Internet platforms such as Weibo and Moments. How can funders quickly obtain the content they need from the text information of donation-based crowdfunding. How sponsors can obtain financial support has become a very urgent need. This article uses deep learning methods to process the titles of donation-based crowdfunding, and realizes the classification of donation-based crowdfunding texts in different language styles. Research has found that the BERT+CNN-based donation-based crowdfunding title classification model can more accurately classify titles, and is superior to other models in various evaluation indicators. The research results have practical significance for the research in the field of text classification.