Journal
INFORMATION SCIENCES
Volume 559, Issue -, Pages 111-129Publisher
ELSEVIER SCIENCE INC
DOI: 10.1016/j.ins.2021.01.052
Keywords
Neural networks; Relevance propagation; Tabular data; Interpretable machine learning; Knowledge discovery; Feature selection
Categories
Funding
- Fundacao de Amparo a Pesquisa do Estado do Rio Grande do Sul (FAPERGS) [19/2551-0001906-8]
- Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq) [311611/2018-4]
- Alexander von Humboldt-Stiftung (AvH) -Germany [BRA 1190826 HFST CAPES-P]
- Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES) -Brazil [DAAD/CAPES PROBRAL 88881.198766/2018-01]
- Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior -Brasil (CAPES) [001]
- NVIDIA Corporation
Ask authors/readers for more resources
The study introduces a relevance aggregation algorithm that combines the relevance computed from multiple samples by a neural network to generate scores for each input feature. Two visualization methods for learned patterns were presented to enhance model comprehension. The method accurately identifies the most important features for network predictions.
The lack of interpretability of neural networks is partially why they are not adopted in a wider variety of applications. Many works focus on explaining their predictions, but few take tabular data into consideration, which led to a small adoption even though this data is of high academic and business interest. We present relevance aggregation, an algorithm that combines the relevance computed from several samples as learned by a neural network and generates scores for each input feature. We also present two methods for visualizing the learned patterns, leading to a better model comprehension. The method was tested in synthetic and real-world datasets (breast cancer gene expression, online shopping behavior, and national high school exam) for classification and regression tasks. It correctly identified which features are the most important for the network's predictions. The selected features can be distinct for each class. The rank of features scores also matches their contribution to the model's performance. The results selected relevant features from the data, paving the way for knowledge discovery. The top-ranked features were consistently able to improve the performance of another independent classifier. For poorly trained neural networks, relevance aggregation helped identify incorrect rules or machine bias. (C) 2021 Elsevier Inc. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available