Sensitivity Analysis for Neural Networks

Yeung, Daniel; Cloete, Ian; Shi, Dongcheng; Ng, Wing W. Y.

doi:10.1007/978-3-642-02532-7

Cited by 63 publications

(42 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data mining methods are routinely used by practitioners in everyday management applications. This paper persuades the readers that consumer behaviour is a crucial aspect of data mining which has remained only a little attention so far, although sensitivity of data mining methods to the presence of outlying observations has been repeatedly reported as a serious problem (Wong, 1997;Yeung et al, 2010). The study has assessed five key consumer behavior factors of respondents namely Carefullness, Shrewdness, Cost Consciousness, Triviality, and Undisturbed.…”

Section: Resultsmentioning

confidence: 99%

Knowledge discovery from consumer behavior in electronic home appliances market in Chennai by using data mining techniques

Vijayalakshmi¹,

Mahalakshmi²,

Magesh³

2013

Afr. J. Bus. Manage.

View full text Add to dashboard Cite

The global economy is improving every year and during the forthcoming decades, marketers need to enter new national markets towards an understanding of how data mining techniques influences consumer behavior, which will be vital for consumer researches. The comprehension of available data mining methods to the presence of outlying measurements in the observed data is discussed as a major drawback of existing data mining methods. The psychological and social processes involved in consumer behaviour forms the subject matter of this study. The objective in accordance with an optimistic approach in terms of studying cause and effect in consumer behaviour will be combined with interpretative prominence on trying to understand the emotional, non-rational aspects of the process. The scope of this paper is to: (1) provide knowledge discovery in consumer behavior, (2) provide experience in the application of K-means data mining techniques in consumer behavior concepts to marketing management decisions. The methodology involves through systematic sampling method and prepared questionnaire which helps to discover knowledge from consumer behaviour predominantly through data mining for the extraction of hidden predictive information from large databases organizations can recognize valuable customers, predict future behaviors, and enable firms to make practical, knowledge-driven decisions. The study will be based on market segmentation wherein, retailers will realize they could no longer sell whatever they bought but had to begin competing for their businesses. This paper proposes k means clustering methods and dendograms suitable for the analysis of data in management applications.

show abstract

Section: Resultsmentioning

confidence: 99%

Knowledge discovery from consumer behavior in electronic home appliances market in Chennai by using data mining techniques

Vijayalakshmi¹,

Mahalakshmi²,

Magesh³

2013

Afr. J. Bus. Manage.

View full text Add to dashboard Cite

show abstract

“…Under conditions when the aggregation signal exceeds a threshold value, the neuron generates an output signal, which is transferred to other neurons. After the signal is transferred, the next weighting processes are started, thus supporting the process of information-processing [11,12]. The mathematical model for the model of artificial neuron is given in Eq.…”

Section: Black-box Modellingmentioning

confidence: 99%

Combined and Mixed Methods Research in Environmental Engineering: When Two is Better Than One

Timma

Blumberga

2015

Energy Procedia

View full text Add to dashboard Cite

Eco-innovation will play a major role in contributing to a sustainable future. Understanding the diffusion is critical, since the diffusion rate is slow and the path unclear. We present the conceptual methodology for eco-innovation diffusion. This methodology combines mixed and combined research methodology to improve the validity and reliability of results. A survey is used as a quantitative method; artificial neural networks and system dynamic are applied as qualitative methods. With this article we aim to contribute to the applications of combined and mixed methods research in the field of eco-innovation diffusion. The application of the presented methodology can be extended for other problems.

show abstract

“…The DBM approach (Young and Beven, 1994) for hydrological model development is of particular relevance as it offers a recognised means by which the legitimacy of a hydrolog- ical model's mechanistic behaviours can be evaluated in the absence of explicit, a priori knowledge about its governing equations. In the DBM approach, a model's mechanistic behaviour is assessed using a formal process of statistical inference through which the required modelling mechanisms and behaviours are identified prior to building the model, and interpreted according to the extent to which they conform to the nature of the system under study (Young et al, 2004) (Fig.…”

Section: The Data-driven Mechanistic Modelling Frameworkmentioning

confidence: 99%

“…Low coherence is indicative of a model that applies a distinctly different modelling mechanism to each local data point and is a means by which data overfitting may be detected. Although methods for computing relative parameter sensitivities are not yet available for all DDMs, recent work has focussed on how it may be achieved for ANN models (Yeung et al, 2010). This has provided new opportunities for exploring their mechanistic behaviour within the DDMMF.…”

Section: Enabling the Ddmmf For Ann Models: Revealing Mechanistic Behmentioning

confidence: 99%

See 1 more Smart Citation

Legitimising data-driven models: exemplification of a new data-driven mechanistic modelling framework

Mount

Dawson

Abrahart

2013

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract.In this paper the difficult problem of how to legitimise data-driven hydrological models is addressed using an example of a simple artificial neural network modelling problem. Many data-driven models in hydrology have been criticised for their black-box characteristics, which prohibit adequate understanding of their mechanistic behaviour and restrict their wider heuristic value. In response, presented here is a new generic data-driven mechanistic modelling framework. The framework is significant because it incorporates an evaluation of the legitimacy of a data-driven model's internal modelling mechanism as a core element in the modelling process. The framework's value is demonstrated by two simple artificial neural network river forecasting scenarios. We develop a novel adaptation of first-order partial derivative, relative sensitivity analysis to enable each model's mechanistic legitimacy to be evaluated within the framework. The results demonstrate the limitations of standard, goodness-of-fit validation procedures by highlighting how the internal mechanisms of complex models that produce the best fit scores can have lower mechanistic legitimacy than simpler counterparts whose scores are only slightly inferior.

show abstract

Sensitivity Analysis for Neural Networks

Cited by 63 publications

References 35 publications

Knowledge discovery from consumer behavior in electronic home appliances market in Chennai by using data mining techniques

Knowledge discovery from consumer behavior in electronic home appliances market in Chennai by using data mining techniques

Combined and Mixed Methods Research in Environmental Engineering: When Two is Better Than One

Legitimising data-driven models: exemplification of a new data-driven mechanistic modelling framework

Contact Info

Product

Resources

About