DBC-Forest: Deep forest with binning confidence screening

TL;DR

DBC-Forest enhances deep confidence screening forests by binning instances based on confidence levels, effectively identifying mis-partitioned instances, leading to faster and more accurate predictions.

Contribution

This paper introduces DBC-Forest, a novel method that improves confidence screening by binning instances, reducing mis-partitioned cases and increasing prediction accuracy and speed.

Findings

Achieves higher accuracy with the same hyperparameters.

Faster than comparable models at the same accuracy.

Effectively identifies mis-partitioned instances.

Abstract

As a deep learning model, deep confidence screening forest (gcForestcs) has achieved great success in various applications. Compared with the traditional deep forest approach, gcForestcs effectively reduces the high time cost by passing some instances in the high-confidence region directly to the final stage. However, there is a group of instances with low accuracy in the high-confidence region, which are called mis-partitioned instances. To find these mis-partitioned instances, this paper proposes a deep binning confidence screening forest (DBC-Forest) model, which packs all instances into bins based on their confidences. In this way, more accurate instances can be passed to the final stage, and the performance is improved. Experimental results show that DBC-Forest achieves highly accurate predictions for the same hyperparameters and is faster than other similar models to achieve the same accuracy.