Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks

By Julian D. Olden, Donald A. Jackson. 2002

1,212 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Environmental Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Environmental Engineering

Citations in the corpus

Data-Based Models for the Prediction of Dam Behaviour: A Review and Some Methodological Considerations (2017, background): “More complex algorithms have been proposed in related fields to unveil the relevance of each input in NN models (see for example [67], [68] and [69]), which may be helpful in dam monitoring.”
A review of artificial neural network models for ambient air pollution prediction (2019, background): “That is, a model is structurally valid if it not only correctly approximates the observed data, but also reflects the way in which the (Mount et al., 2013), overall connection weights (Olden & Jackson, 2002), and measure of generalisation (Razavi & Tolson, 2011).”
Generalized models vs. classification tree analysis: Predicting spatial distributions of plant species at different scales (2003, methodology): “…1991), regression and classification tree analyses (RTA and CTA) (Breiman et al. 1984) and other rule-based methods such as neural networks (Olden & Jackson 2002; Pearson et al. 2002) or cellular automata (Carey 1996), for which no assumptions are made regarding species response curves to…”
Opening the black box - data driven visualization of neural networks (2005, methodology): “Olden and Jackson [21] introduce a randomization approach to statistically analyze the input importance based on Garson’s method [12].”
A review of statistical methods for the evaluation of aquatic habitat suitability for instream flow assessment (2006, methodology): “ANNs are receiving increased attention in the ecological studies as a powerful modelling paradigm for uncovering patterns in ecological data (Olden and Jackson, 2002a).”
Assessing transferability of ecological models: an underappreciated aspect of statistical validation (2012, background): “Neural networks offer multiple ways to control complexity, including managing the number of nodes (as we did here), stopping the algorithm early and weight decay or weight elimination (Bishop 1995; Sarle 1995; Olden & Jackson 2002).”
Can machine-learning improve cardiovascular risk prediction using routine clinical data? (2017, background): “However, improvements in data visualization methods have improved understanding of these models, illustrating the importance of network connections between risk factors [35] (See example visualising our neural network model in Fig 2).”
Anxiety, panic and self-optimization (2018, methodology): “Recommended videos are selected by YouTube’s ‘black-boxed’ algorithm, so called because YouTube ‘provides little explanatory insight into the relative influence of the independent variables in the process’ (Olden and Jackson, 2002: 135). and researchers (Kitchin, 2016).”
A Comparative Study of PSO-ANN, GA-ANN, ICA-ANN, and ABC-ANN in Estimating the Heating Load of Buildings’ Energy Efficiency for Smart City Planning (2019, methodology): “This method enables the analysis of the importance of input variables for hidden multiple-layer ANN models [79].”
The use of artificial neural networks to analyze and predict alongshore sediment transport (2010, background): “However, for more than a decade techniques have been suggested that allow detailed analysis of connection weights and estimation of the role of each input variable (Vaughn, 1996; Benı́t z, 1997; Dimopoulos et al., 1999; Olden, 2000; Olden and Jackson, 2002).”
Protocol for developing ANN models and its application to the assessment of the quality of the ANN model development process in drinking water quality modelling (2014, methodology): “…of an ANN model can be assessed using a variety of techniques, such as sensitivity analysis (Mount et al., 2013), overall connection weights (Olden and Jackson, 2002; Kingston et al., 2005a, 2006), a measure of generalization (Razavi and Tolson, 2011) and other methods, such as comparing…”