Risk assessment of human neural tube defects using a Bayesian belief network

Abstract

Neural tube defects (NTDs) constitute the most common type of birth defects. How much risk of NTDs could an area take? The answer to this question will help people understand the geographical distribution of NTDs and explore its environmental causes. Most existing methods usually take the spatial correlation of cases into account and rarely consider the effect of environmental factors. However, especially in rural areas, the NTDs cases have a little effect on each other across space, whereas the role of environmental factors is significant. To demonstrate these points, Heshun, a county with the highest rate of NTDs in China, was selected as the region of interest in the study. Bayesian belief network was used to quantify the probability of NTDs occurred at villages with no births. The study indicated that the proposed method was easy to apply and high accuracy was achieved at a 95% confidence level.

Optimal discretization for geographical detectors-based risk assessment

“…Spatial data mining methods such as neural networks, support vector machines, rough sets, and Bayesian networks mine the potential associations between health and its risk factors. They can also provide effective health risk assessments and predictions (Lisboa and Taktak 2006;Magnin 2009;Bai et al 2010;Liao et al 2010). The geographical detectors model is a new spatial analysis method used to assess health and environmental risks.…”

Bi-directional risk assessment in carbon capture and storage with Bayesian Networks

“…BNs, also called causal or probabilistic networks, are largely developed by the artificial intelligence community and they have been applied in a number of diverse problem domains including ecological modelling (Uusitalo, 2007), medical diagnosis (Wiegerinck et al, 1999), image classification (Malka and Lerner, 2004) and fraud detection (Kirkos et al, 2007). A number of publications have used BNs as a method of addressing risk assessment, such as in nuclear waste disposal (Lee and Lee, 2006), neural tube defects (Liao et al, 2010) and in seismic risk (Bayraktarli et al, 2006). Recently BNs have started to make their way into the CCS community for analysing safety risk related to loss of containment in CO 2 transport (Kvien et al, 2013), for combining evidence from multiple CO 2 leak detection technologies in geological storage (Yang et al, 2012) and for discriminating between natural, triggered and induced earthquakes in areas that have geo-engineering operations (Dahm et al, 2010).…”

Applying Bayesian belief networks to health risk assessment

“…The objectives of these papers can be roughly divided into four groups: prediction, evaluation, diagnosis and classification. Examples of prediction cases were the work of Dlamini (2010) who used the BBN to predict wildfire occurrence in Swaziland; and the study of Liao et al (2010) who used BBN to predict the rate of human neural tube defects by considering the number of doctors, the use of pesticides and fertilizer, the production of vegetable and fruit, percapita net incomes, elevation, NDVI, road and fault buffer, influence of coal mines and distances to the nearest factory. Example studies of evaluation were the work of Ticehurst et al (2010) who used the BBN to complement conventional analyses for exploring landholder management of native vegetation; and the paper of Ordóñez Galán et al (2009) who used the BBN to suggest a reforestation project for woodland types based on existing woodland types in the area and environmental variables such as altitude, slope, potential insolation, lithology, rainfall and distance from the sea.…”