Decision Tree Based Crowd Funding for Kickstarter Projects

The proposed work employs the C4.5 decision tree algorithm on a kick-starter project dataset to help a user decide whether to back a kick-starter project that is ongoing by predicting how likely it is that it may be a successful one. We pre-processed the kick-starter dataset with about 35 columns, and used WEKA to run the algorithm on the dataset. We reached an accuracy of 99.7% and we also talk about why the algorithm chose 5 particular attributes over the others. A lot of other papers have discussed this problem from a project creator's standpoint, predicting whether a project is going to be a success before it has begun. There are fewer papers which look into predicting the success of the ongoing projects that helps users choose potentially successful projects to back, and we have also achieved a higher accuracy rate.