Novel rough neural network for classification with missing data

The paper presents a new feedforward neural network architecture. Thanks to incorporating the rough set theory, the new network is able to process imperfect input data, i.e. in the form of intervals or with missing values. The paper focuses on the last case. In contrast to imputation, marginalisation and similar solutions, the proposed architecture is able to give an imprecise answer as the result of input data imperfection. In the extreme case, the answer can be indefinite contrary to a confabulation specific for the aforementioned methods. The results of experiments performed on three classification benchmark datasets for every possible combination of missing attribute values showed the proposed solution works well with missing data with accuracy dependent on the level of missing data.

[1]  Nicola Sartori,et al.  Multiple imputation of missing values in a cancer mortality analysis with estimated exposure dose , 2005, Comput. Stat. Data Anal..

[2]  Marcin Korytkowski,et al.  Ensemble of rough-neuro-fuzzy systems for classification with missing features , 2008, 2008 IEEE International Conference on Fuzzy Systems (IEEE World Congress on Computational Intelligence).

[3]  Robert Nowicki,et al.  On Combining Neuro-Fuzzy Architectures with the Rough Set Theory to Solve Classification Problems with Incomplete Data , 2008, IEEE Transactions on Knowledge and Data Engineering.

[4]  Christian Napoli,et al.  Enhancing Environmental Surveillance Against Organised Crime with Radial Basis Neural Networks , 2015, 2015 IEEE Symposium Series on Computational Intelligence.

[5]  Robert Nowicki,et al.  Rough Neuro-Fuzzy Structures for Classification With Missing Data , 2009, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[6]  T. Stijnen,et al.  Review: a gentle introduction to imputation of missing values. , 2006, Journal of clinical epidemiology.

[7]  Volker Tresp,et al.  Training Neural Networks with Deficient Data , 1993, NIPS.

[8]  Aleksandra Rutkowska.  Influence Of Membership Function’s Shape On Portfolio Optimization Results , 2016, J. Artif. Intell. Soft Comput. Res..

[9]  Marcin Woźniak,et al.  Rough k nearest neighbours for classification in the case of missing input data , 2014 .

[10]  D. Rubin.  INFERENCE AND MISSING DATA , 1975 .

[11]  Robert Nowicki,et al.  On classification with missing data using rough-neuro-fuzzy systems , 2010, Int. J. Appl. Math. Comput. Sci..

[12]  Eren Bas,et al.  The Training Of Multiplicative Neuron Model Based Artificial Neural Networks With Differential Evolution Algorithm For Forecasting , 2016, J. Artif. Intell. Soft Comput. Res..

[13]  Robert Nowicki,et al.  Rough Deep Belief Network - Application to Incomplete Handwritten Digits Pattern Classification , 2015, ICIST.

[14]  Marcin Korytkowski,et al.  AdaBoost Ensemble of DCOG Rough-Neuro-Fuzzy Systems , 2011, ICCCI.

[15]  Volker Tresp,et al.  Efficient Methods for Dealing with Missing Data in Supervised Learning , 1994, NIPS.

[16]  Marcin Woźniak,et al.  Automatic classification of fruit defects based on co-occurrence matrix and neural networks , 2015, 2015 Federated Conference on Computer Science and Information Systems (FedCSIS).

[17]  K. Anitha,et al.  Rough Neural Network , 2016 .

[18]  Amit Gupta,et al.  Estimating Missing Values Using Neural Networks , 1996 .

[19]  M Daszykowski,et al.  Dealing with missing values and outliers in principal component analysis. , 2007, Talanta.

[20]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[21]  Marcin Gabryel,et al.  Recursive Probabilistic Neural Networks , 2003, PPAM.

[22]  Ryotaro Kamimura,et al.  Accumulative Information Enhancement In The Self-Organizing Maps And Its Application To The Analysis Of Mission Statements , 2015, J. Artif. Intell. Soft Comput. Res..

[23]  Pawan Lingras,et al.  Comparison of Neofuzzy and Rough Neural Networks , 1998, Inf. Sci..