Choosing a linear regression trainer

When looking at the list of nine linear regression trainers ML.NET, it can be a bit daunting to ask which is the best.

For ML.NET linear regression trainers, by and large, the most popular are FastTree and LightGBM. The three FastTree algorithms utilize neighbor-joining and use heuristics to quickly identify candidate joins to build out a decision tree. LightGBM is a very popular linear regression algorithm that utilizes a Gradient-based One Side Sampling (GOSS) to filter out the data instances for finding a split value. Both trainers provide both quick training and predict times while also providing very accurate model performance. Also, more documentation, papers, and research are available with both of these algorithms.