A novel Random Forest integrated model for imbalanced data classification problem

Highlights：

• A novel classification algorithm is proposed based on the integrated over-sampling technique for various imbalanced data sets.

• The new instances are generated dynamically and variously according to the number of the majority class and the sample center of the minority class.

• The improved SSA was adopted to optimize RF parameters adaptively for adapting to each imbalanced data set.

• Most recently used assessment parameters and optimized algorithms for imbalanced data sets classification were diversely and critically compared.