A Boolean Algebra Based Rule Extraction Algorithm for Neural Networks with Binary or Bipolar Inputs

布林代數爲基的法則擷取演算法以運用於二進制或二極輸入的類神經網路

楊烽正(Feng-Cheng Yang) ； 黃俊凱(Chun-Kai Hwang)

《工業工程學刊》 21卷1期 (2004/01) Pp. 27-39

https://doi.org/10.29977/JCIIE.200401.0003

摘要

類神經網路被應用在各種領域，包括科學、商務、醫學、以及各種產業上。但是隱藏在經過週延成熟訓練的類神經網路內的知識難以讓人瞭解。本文提出一種布林代數為基的演算法，從監督、前饋式的類神經網路中擷取可以理解的布林規則，以揭露黑箱內的知識。這種演算法稱為BAB-BB規則擷取演算法，代表布林代數為基的規則擷取演算法應用於二進制和二極輸入的類神經網路。解構技術和區間代數是演算法使用的主要技術。首先，分析各個神經元和它的輸入，以建立一條描述因它的輸入激發出該神經元輸出間的的布林函數。這些神經元的布林函數根據網路拓撲結構，合併成一個被累加的布林規則。這個繁複的布林規則再進一步由布林代數的各種化簡法則予以簡化。在規則擷取過程中，可以檢算出多餘無效的隱藏層神經元。這些神經元可以逕行刪除且不會影響類神經網路的原始功能。文中也提出多個二進制和二極的範例輸入資料，以展示本演算法的功能。此外，本文也使用本演算法擷取執行布林互斥或運算的神經網路內的法則。結果表示，我們的BAB-BB演算法確實可行並有適度的求解效率。

關鍵字

類神經網路 ； 法則擷取 ； 知識萃取 ； 資料探擷

並列摘要

Neural networks have been applied in various domain including science, commerce, medicine, and industry. However, The knowledge learned by a trained neural network is difficult to understand. This paper proposes a Boolean algebra based algorithm to extract comprehensible Boolean rules from supervised feed-forward neural networks to uncover the black-boxed knowledge. This algorithm is called the BAB-BB rule extraction algorithm, which stands for a Boolean algebra based rule extraction algorithm for neural networks with binary and bipolar inputs. Decomposition techniques and interval arithmetic are used in the algorithm. First, each neuron associated with its inputs is analyzed and a Boolean function, describing the activation rule from its inputs to the neuron, is derived. These Boolean functions are merged into an aggregated Boolean rule according to the network topology. The Boolean rule is then further simplified by Boolean algebra operations. During the rule extraction procedure, redundant hidden neurons can be detected and removed without affecting the original function of the neural network. Examples of unipolar and bipolar inputs are presented to demonstrate the use of our algorithm. Finally, the Exclusive OR problem is presented and solved by our algorithm. Results show that our BAB-BB algorithm is practicable and of high efficiency.

並列關鍵字

neural networks ； rule extraction ； knowledge discovery ； data mining

參考文獻

Setiono, R.,J. Thong, L. Y,C. S. Yap(1998).Case study: symbolic rule extraction from neural networks an application to identifying organizations adopting IT.Information and Management.34,91-101.

Zhou, Y.,Y. Lu,C. Shi(1997).Using neural network to extract knowledge from database.Proceedings of The First European Symposium, PKDD.97,376-383.

McGarry, K. J.,J. Tait, S. Wermter,MacIntyre, J.(1999).Rule-extraction from radial basis function networks.Ninth International Conference on Artificial Neural Networks, ICANN 99.99(2),613-618.

Andrews, R.,J. Diederich,A. B. Tickle(1995).Survey and critique of techniques for extracting rules from trained artificial neural networks.Knowledge-Based Systems.8(6),373-389.

Haykin, S.(1999).Neural Networks: A Comprehensive Foundation.