Inspiration
Speaking with Josiah about the goal of Marky inspired us to join in on the cause to help smaller brands break into the social media industry to grow their company. However, we mutually recognized the existing weaknesses of modern AI, so it only seems natural to discover how variations in posts can lead to an approved post!
What it does
Our algorithm analyzes trends in both tabular data and graphic styling to uncover the key features that impact the probability of getting approved.
How we built it
In order to do tabular classification, we used XGBoost, an extreme gradient boosting algorithm, to analyze the tabular data. Then, we used convolutional neural networks to uncover key features in the image.
What makes a good post -> A combination of both graphics and text content. All of it should meld together in order to serve a purpose whether it is to inform, convince, inspire...etc.
Log in or sign up for Devpost to join the conversation.