![]() ![]() High score means the attribute contributed positively to the the final decision (Funded: yes), while low scores contributed negatively. Lastly, not having a Facebook page or a video amounts to almost nothing in the making of the final prediction. On the other side, we see how the duration of the project, description length, maximal pledge tiers and the type of the idea work against the decision to fund the project. ![]() The fact that there were 11 pledge levels, 13 images, many connections to other projects and the length of the project description - all of these attributes add something positive to the funding. The highest ranking attributes are those that contributed the most (high Score value). Finally, select Create Games & Apps Without Any Coding in the Data Table and connect it to the widget. Then add the classifier, say, Logistic Regression. Connect File widget with Explain Predictions. Our data set, a classifier and a data sample that we wish to inspect. Now, let’s see why the app Create Games & Apps Without Any Coding got funded.Įxplain Predictions needs 3 inputs. ![]() Select the data instance you wish to explore in a Data Table. Or at least, explain why the classifier thinks they will.įirst, let us load the Kickstarter data from the Datasets widget and inspect it in a Data Table. But why is that? When we are looking for possible explanations, it is easy to ascribe the failure to the type of the idea.īut what about those rare cases, where an app idea gets funded? Can we figure out why a particular idea succeeded? Our new widget Explain Predictions can do just that - explain why they will succeed. On Kickstarter most app ideas don’t get funded. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |