FERNN: An Algorithm for Fast Extraction of Rules from Neural Networks

By Rudy Setiono, Wee Kheng Leow. 2004

155 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Tameru Hailesilassie
2. Thangairulappan Kathirvalavakumar
3. Gethsiyal Augasta M
4. Leilani H. Gilpin
5. Lalana Kagal
6. David Bau
7. Ben Z. Yuan
8. Michael A. Specter

In Computer Vision and Pattern Recognition

1. Tameru Hailesilassie
2. Novruz M. Allahverdi
3. Thangairulappan Kathirvalavakumar
4. Humar Kahramanli
5. Gethsiyal Augasta M
6. S. Kotsiantis
7. Leilani H. Gilpin
8. Ben Z. Yuan

Citations in the corpus

• Explaining Explanations: An Overview of Interpretability of Machine Learning (2018, methodology): “Similarly, The FERNN [37] algorithm uses the C4.5 algorithm [24] and tries to identify the meaningful hidden neurons and inputs to a particular network.”
• Rule Extraction Algorithm for Deep Neural Networks: A Review (2016, methodology): “Even though this algorithm is independent of network structure and training algorithm, it is not possible to apply the algorithm directly to DNN. FERNN [18] is an algorithm proposed to generate both M-of-N and IF-THEN rule from feedforward neural network with a single hidden layer.”
• Rule extraction from trained adaptive neural networks using artificial immune systems (2009, methodology): “Setiono and Leow (2000) presented the fast method is based on the relevance of hidden units, considering their information gains.”
• Extracting symbolic rules from trained neural network ensembles (2003, background): “Setiono and Leow [37] divide the activation values of relevant hidden units to two subintervals and then find out the set of relevant connections to those relevant units to construct rules.”
• Development and Application of Artificial Neural Network (2017, background): “The hidden unit is a unit that is located between the input and output units and cannot be observed outside the system [6].”
• Explaining Explanations: An Approach to Evaluating Interpretability of Machine Learning (2018, methodology): “Similarly, The FERNN [37] algorithm uses the C4.5 algorithm [24] and tries to identify the meaningful hidden neurons and inputs to a particular network.”
• Computational intelligence methods for rule-based data understanding (2004, methodology): “Several neural algorithms for rule extraction have been compared experimentally on benchmark datasets with very good results [4], [11], [57]–[60].”
• Rule extraction from neural networks — A comparative study (2012, methodology): “FERNN [17] is a decompositional algorithm for classification rule extraction from feedforward neural networks.”
• Supervised Machine Learning: A Review of Classification Techniques (2007, background): “Setiono and Leow (2000) divided the activation values of relevant hidden units into two subintervals and then found the set of relevant connections of those relevant units to construct rules.”