Data Augmentation Using Gaussian Mixture Model on CSV Files

摘要: One of the biggest challenges in training supervised models is lack amount labeled data for model and facing overfitting underfitting problems. solutions solving this problem augmentation. There have been many developments augmentation image files, especially medical type datasets, by doing some changes on original file such as Random cropping, Filliping, Rotating, so on, order to make a new sample file. Or use Deep Learning generate similar samples like Generative Adversarial Networks, Convolutional Neural Networks on. However, numerical dataset, there not enough advances. In paper, we are proposing Gaussian Mixture Models (GMMs) augment more very Numerical dataset. The results demonstrated that Mean Absolute Error decreases meaning regression became accurate.