Improve Er Method By Machine Learning Methods In Factor Model With Strong Individual Effects

This paper focus on deep comprehension of factor model with strong individual effects and ER method.Try to find machine learning improve methods when ER has bad expression.Compare with monotone bounded transformation and has better expression which can not only estimate the true factor number but can reflect the strength of individual effects of model.The statement not only provide methods but also describe in the order of thinking of solve problem.Focus on the Algorithms' implement and optimization.