Abstract:A novel multi-channel ensemble framework is proposed for intelligent fault diagnosis with imbalance data.A hybrid sampling method is proposed for alleviate the imbalance of training data and providing rebalanced data for each cascade forest channel.A new type of balanced forest is proposed for improving the accuracy of the basic classifier in the ensemble method.The effectiveness of the method is tested by experimental measurements.Data imbalance is a common problem in rotating machinery fault diagnosis. Traditional data-driven diagnosis methods, which learn fault features based on balance data set, would be significantly affected by imbalanced data. In this paper, a novel imbalanced data related fault diagnosis method named deep balanced cascade forest is proposed to solve this problem. deep balanced cascade forest is a multi-channel cascade forest, in which, each of its channels adaptively generates deep cascade structure and is trained on independent data. To enhance the performance of imbalance classification, the deep balanced cascade forest is promoted from both aspects of resampling and algorithm design. A hybrid sampling method, namely Up-down Sampling, is proposed to provide rebalanced data for each cascade forest channel. Meanwhile, a new type of balanced forest with an improved balanced information entropy for attribute selection is designed as the basic classifier of cascade forest. The good synergy of these two methods is the key to the deep balanced cascade forest model. This good synergy makes deep balanced cascade forest achieve the fusion of data-level methods and algorithm-level methods. Comparative experiments on sufficient imbalanced data sets have been designed to verify the performance of the proposed model, and results confirm that deep balanced cascade forest is much more stable and effective in handling imbalance fault diagnosis problem compared to the popular deep learning methods.

WCDForest: a Weighted Cascade Deep Forest Model Toward the Classification Tasks

Very High Resolution Remote Sensing Imagery Classification Using a Fusion of Random Forest and Deep Learning Technique—Subtropical Area for Example

An improved deep forest for alleviating the data imbalance problem

Real-time Semantic Segmentation with Weighted Factorized-Depthwise Convolution

Improving deep forest by ensemble pruning based on feature vectorization and quantum walks

TSCF: An Improved Deep Forest Model for Time Series Classification

Improving Deep Forest by Confidence Screening.

Deep Fuzzy Rule-based Classification System with Improved Wang-Mendel Method

Deep Forest with Hashing Screening and Window Screening

Interpreting Deep Forest through Feature Contribution and MDI Feature Importance

Deep Dynamic Boosted Forest.

A New Rotation Forest Ensemble Algorithm

Improving Deep Forest Via Patch-Based Pooling, Morphological Profiling, and Pseudo Labeling for Remote Sensing Image Classification

Improve Deep Forest with Learnable Layerwise Augmentation Policy Schedule

Deep Forest: Towards an Alternative to Deep Neural Networks

Deep Forest: Towards an Alternative to Deep Neural Networks.

ForestLayer: Efficient Training of Deep Forests on Distributed Task-Parallel Platforms

Deep balanced cascade forest: An novel fault diagnosis method for data imbalance

Improving Deep Forest by Exploiting High-order Interactions

Deep ensemble forests for industrial fault classification