- The project is based on Kaggle Competition on Social Influencers. It is availabe here : https://www.kaggle.com/c/predict-who-is-more-influential-in-a-social-network.
- This is a binary-class text classification problem.
- We tried several models : Logistic Regression, Gaussian Naive Bayes, Neural Nets, Boosting, SVM.
The following are the results on the test dataset. The results represent Area under the ROC curve.
Model | AUC |
---|---|
Logistic Regression | 0.8606 |
XgBoost | 0.86168 |
Gaussian Naive Bayes | 0.82009 |
Neural Nets | 0.8590 |
SVM | 0.8376 |