While more and more people are engaged in crowdfunding activities, we know little about the quantitative side of crowdfunding: what type of campaigns are successfully raising funding and what types are not?
In the talk, I will present a machine learning model on whether a crowdfunding campaign is successful or not. The focus of the model is on text features that I create from the project descriptions and the titles and test whether tl;dr might affect the backers to donate money. I correlate these features with the success of the campaign. Moreover, I will create image features from pictures that are uploaded together with the project description and other information about the crowdfunding campaigns. My analysis is not related to real Artificial Intelligence, but I use weights from a pre-trained deep learning model to extract image features. I will show how „surprisingly” well this pre-trained model describes unknown images.
At the end of the talk we will have shed some light on whether or not tl;dr and crowdfunding success are related to each other. You might even be inspired to use some free tools to estimate your own model : )