# Introduction We briefly introduced the data processing methods of linear distribution and non-linear distribution, and presented some actual experimental operations. There is also another machine learning method to solve small samples: non-linear and high-dimensional pattern recognition that has shown many unique advantages. Especially in the case of small sample size, its practical application even outperforms neural networks. Moreover, it can be applied not only to linearly distributed data, but also to non-linearly distributed data. Compared to other basic machine learning classification algorithms such as logistic regression, KNN, naive Bayes etc., it generally performs far better.
Click the virtual machine below to start practicing