Soft Margin MaximizationIt is an optimization method that mainly uses soft intervals to find the optimal solution. Selecting the best separating hyperplane is an optimization problem, which is based on how to maximize the geometric interval.

However, in practice, to maximize the geometric interval of the training data set, it is necessary that the geometric interval of the classification hyperplane of all training samples is greater than this value. Different from hard interval maximization, soft interval maximization adds the concept of slack variables. By using soft intervals to divide the hyperplane, the influence of outliers on the optimal classification hyperplane can be eliminated.

Related terms: soft margin, hard margin maximization, classification problem