On Neural Network Architecture Based on Concept Lattices
作者： Sergei O. Kuznetsov , Nurtas Makhazhanov , Maxim Ushakov
DOI: 10.1007/978-3-319-60438-1_64
关键词:
摘要: Selecting an appropriate network architecture is a crucial problem when looking for solution based on neural network. If the number of neurons in too high, then it likely to overfit. Neural networks also suffer from poor interpretability learning results. In this paper approach building concept lattices and coming monotone Galois connections proposed attempt overcome mentioned difficulties.
springer.com
sci-hub.st 参考文章(15)
Bernhard Ganter, Sergei O. Kuznetsov, Hypotheses and Version Spaces international conference on conceptual structures. pp. 83- 95 ,(2003) , 10.1007/978-3-540-45091-7_6
Engelbert Mephu Nguifo, Norbert Tsopze, Gilbert Tindo, M-CLANN: Multiclass Concept Lattice-Based Artificial Neural Network Constructive Neural Networks. pp. 103- 121 ,(2009) , 10.1007/978-3-642-04512-7_6
Sergei O. Kuznetsov, Fitting Pattern Structures to Knowledge Discovery in Big Data international conference on formal concept analysis. pp. 254- 266 ,(2013) , 10.1007/978-3-642-38317-5_17
Bernhard Ganter, Rudolf Wille, C. Franzke, Formal Concept Analysis: Mathematical Foundations ,(1998)
Sergei O. Kuznetsov, Machine Learning and Formal Concept Analysis Concept Lattices. pp. 287- 312 ,(2004) , 10.1007/978-3-540-24651-0_25
Dean van der Merwe, Sergei Obiedkov, Derrick Kourie, AddIntent: A New Incremental Algorithm for Constructing Concept Lattices Concept Lattices. pp. 372- 385 ,(2004) , 10.1007/978-3-540-24651-0_31
Bernhard Ganter, Sergei O. Kuznetsov, Formalizing Hypotheses with Concepts international conference on conceptual structures. pp. 342- 356 ,(2000) , 10.1007/10722280_24
Sebastian Rudolph, Using FCA for Encoding Closure Operators into Neural Networks international conference on conceptual structures. pp. 321- 332 ,(2007) , 10.1007/978-3-540-73681-3_24
Geoffrey G. Towell, Jude W. Shavlik, Knowledge-based artificial neural networks Artificial Intelligence. ,vol. 70, pp. 119- 165 ,(1994) , 10.1016/0004-3702(94)90105-8
Renato Vimieiro, Pablo Moscato, Disclosed: An efficient depth-first, top-down algorithm for mining disjunctive closed itemsets in high-dimensional data Information Sciences. ,vol. 280, pp. 171- 187 ,(2014) , 10.1016/J.INS.2014.04.044