Inspiration
Gamers on Mistplay are rewarded for playing, engaging and buying stuff in mobile games. All these games are freemium, where a user can make in-app purchases to unlock levels or buy extra features within the game. We want to build a model to predict if a user will be spending money on a specific game or not.
How we built it
We used SQL to preprocess data, and then used machine learning SVM model to predict the results. We built an interface using C#, HTML and bootstrap to take in new test files and output prediction results.
Challenges we ran into
- Large data size, biased and missing data
- Prediction score(Accuracy) is not a true representation of the result. For the choice of the metrics, F1 score is a more accurate.
- Since we use c# to build the interface, we have trouble running python script on the interface.
Notes
GitHub Link: https://github.com/S1o3urn/CodeJam2019-02
Log in or sign up for Devpost to join the conversation.