Categorical Embeddings for Tabular Data using PyTorch

Deep learning has received much attention for computer vision and natural language processing, but less tabular data, which is the most prevalent type of data used in industry. Embeddings offer a solution by representing categorical variables as continuous vectors lowdimensional space. PyTorch provides excellent support GPU acceleration pre-built functions modules, making it easier to work with embeddings variables. In this research paper, we apply feedforward neural network model multiclass classification problem using Shelter Animal Outcome dataset. We calculate probability an animal's outcome belonging each 5 categories. Additionally, explore feature importance two common techniques: MDI permutation. Understanding crucial building better models, improving performance, interpreting communicating results. Our findings demonstrate usefulness deep highlight selection effective machine models.