We have looked at different ways to split our datasets to build our evaluation strategy. In most cases, the data that we receive may not be in a format that can be readily used by us for training our algorithms. In this section, we will cover some of the preprocessing techniques and feature engineering techniques. Though most of the feature engineering techniques are domain-specific, particularly in the areas of computer vision and text, there are some common feature engineering techniques that are common across the board, which we will discuss in this chapter.
Data preprocessing for neural networks is a process in which we make the data more suitable for the deep learning algorithms to train on. The following are some of the commonly-used data preprocessing steps:
- Vectorization
- Normalization
- Missing values
- Feature extraction
