Density estimation is the basic and common problem in statistics and machine learning estimation. Generally, Density estimation is the technique for the computation of probability density function by learning the relations among attributes in the data. Density estimation provides the skewness and multi-modality in the data for informal investigation. Density estimation approaches are broadly classified into two groups: parametric density estimation and non-parametric density estimation. Parametric probability density estimation involves the investigation of density function from the data sample based on common distribution.
Non-parametric probability density estimation involves a technique that fits a model to the arbitrary distribution of the data samples for density estimation. Traditional density estimation methods such as histogram and kernel density function perform well on low dimensional density estimation problems. Emerge of neural network-based density estimation is to deal with high-dimensional data problems. The approaches for neural density estimation are autoregressive models and normalizing flows. Recent development in density estimation for high dimensional data problems such as Generative adversarial networks.