Bias and variance of validation methods for function approximation neural networks under conditions of sparse data

Twomey, Janet M.; Smith, Alice E.

doi:10.1109/5326.704579

Cited by 105 publications

(51 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The bootstrap error estimate consists of the training error for the application model plus the difference, averaged over all bootstrap models, between the error for the full data sample and the error for the bootstrap sample. Bootstrap appears to produce reliable model validation without an extensive computational effort, and is especially useful if the amount of data available for model construction and validation is limited [11] and when, as in the present study, some of the target cases are under-represented in the number of available examples.…”

Section: Model Validationmentioning

confidence: 93%

Glaucoma detection and evaluation through pattern recognition in standard automated perimetry data

Wróblewski

Francis

Chopra

et al. 2009

Graefes Arch Clin Exp Ophthalmol

View full text Add to dashboard Cite

show abstract

Section: Model Validationmentioning

confidence: 93%

Glaucoma detection and evaluation through pattern recognition in standard automated perimetry data

Wróblewski

Francis

Chopra

et al. 2009

Graefes Arch Clin Exp Ophthalmol

View full text Add to dashboard Cite

show abstract

“…To add further confidence in the validity of the results from this analysis, a re-sampling procedure is undertaken and the models recalculated (see for example [12]). Due to page limitation this validation exercise is undertaken on the incomplete data set only.…”

Section: Validation Analysis Of Ncarbs Results (Using Re-sampling)mentioning

confidence: 99%

Evidence-Based Modelling of Organizational Social Capital with Incomplete Data: An NCaRBS Analysis

Beynon

Andrews

2014

Belief Functions: Theory and Applications

View full text Add to dashboard Cite

Abstract. Organizational social capital is critical to effective organizational functioning. Yet, different aspects of social capital are likely to be present to varying degrees within any given organization. In this study, alternative blends of structural, relational and cognitive social capital are modelled using a range of key organizational variables drawn from an incomplete dataset. A novel evidence-based approach to the ambiguous classification of objects (N-state Classification and Ranking Belief Simplex or NCaRBS) is used for the analysis. NCaRBS is uniquely able to capture the full range of ambiguity in the antecedents and effects of social capital, and to do so by incorporating incomplete data without recourse to the external management of the missing values. The study therefore illustrates the multi-faceted potential of analytical techniques based on uncertain reasoning, using the Dempster-Shafer theory of evidence methodology.

show abstract

“…The output of ANN models was represented in terms of bootstrap resamples and corresponding optimized weights as f NN (x i , w s ) where x i was the input data pattern, and w s was the optimized weights of the ANN model for a particular bootstrap resample s. The performance of both the models was then evaluated using a set A s . Then the generalization error denoted as Ê 0 was estimated (e.g., ANN model) as [TWOMEY, SMITH 1998]: [ ] …”

Section: Bootstrap Techniquementioning

confidence: 99%

Water demand forecasting using extreme learning machines

Tiwari¹,

Adamowski²,

Adamowski³

2016

Journal of Water and Land Development

View full text Add to dashboard Cite

The capacity of recently-developed extreme learning machine (ELM) modelling approaches in forecasting daily urban water demand from limited data, alone or in concert with wavelet analysis (W) or bootstrap (B) methods (i.e., ELM, ELM W , ELM B ), was assessed, and compared to that of equivalent traditional artificial neural network-based models (i.e., ANN, ANN W , ANN B ). The urban water demand forecasting models were developed using 3-year water demand and climate datasets for the city of Calgary, Alberta, Canada. While the hybrid ELM B and ANN B models provided satisfactory 1-day lead-time forecasts of similar accuracy, the ANN W and ELM W models provided greater accuracy, with the ELM W model outperforming the ANN W model. Significant improvement in peak urban water demand prediction was only achieved with the ELM W model. The superiority of the ELM W model over both the ANN W or ANN B models demonstrated the significant role of wavelet transformation in improving the overall performance of the urban water demand model.

show abstract

Bias and variance of validation methods for function approximation neural networks under conditions of sparse data

Cited by 105 publications

References 25 publications

Glaucoma detection and evaluation through pattern recognition in standard automated perimetry data

Glaucoma detection and evaluation through pattern recognition in standard automated perimetry data

Evidence-Based Modelling of Organizational Social Capital with Incomplete Data: An NCaRBS Analysis

Water demand forecasting using extreme learning machines

Contact Info

Product

Resources

About