A Novel Evolutionary Algorithm
nomic status is commonly linked to risk of infestation in the literature ( Briceño-León & Méndez Galván , 2007 ; Prata , 2001 ), these factors are not identified as important features when using other statistical methodologies ( Bustamante et al ., 2014 , 2015 ). Of the three socioeconomic risk factors identified by the CCEA as important , homeownership is associated with many of the archived conjunctive clauses across all three datasets . This feature may be particularly interesting for strategies designed to combat T . dimidiata infestation as the relationship between ownership and high infestation may be viewed several ways . First , it may be easier to persuade the large number of homeowners who live in infested homes to improve their homes ( e . g ., via the adoption of Ecohealth interventions ) because they have the freedom to make improvements should resources be made available ( compared , say , with those renting or borrowing the home ). On the flipside , there may be a connection between economic status poverty and homeownership ; and the reason that infested houses are associated with homeowners is that they cannot afford and / or lack the time to improve their house .
The CCEA is just one of many statistical tools that can be employed to analyze large , complex datasets . The goal of the CCEA is to identify probabilistically significant multivariate interactions that would either ( 1 ) be too costly ( i . e ., from a computational viewpoint ) to search for higher-order , multivariate interactions or ( 2 ) not be detected using additive models . While we use the archived conjunctive clauses to identify important features in this article , these sets of features could be further analyzed in a number of ways . For example , future work may include analyzing infested houses associated with important features ( sets of important features ) for spatial patterns . Sensitivity analysis could be applied to higher-order feature sets to examine whether the model is more likely to be additive or epistatic to help improve mitigation strategies . For example , if both features in an archived second-order conjunctive clause ( e . g ., a large number of chickens and the proximity of the coop to the house ) do not exhibit significant main effects , then the second-order clause exhibits epistatic interactions ( both need to present to be associated with infestation ). However , if both features exhibit significant main effects and when combined the set of features results in a more-fit second-order conjunctive clause , then the second-order conjunctive clause is additive . In addition , archived conjunctive clauses with lots of missing data could be isolated to see if those features show important trends ; and if trends appear important , then more effort might be directed toward ensuring those feature values are collected in future surveys .
The goal of the CCEA algorithm is to provide an archive of conjunctive clauses with strong statistical signals that stakeholders can then use to fulfill their objectives . For instance , if researchers are interested in improving the floor material , they could mine the archived conjunctive clauses associated with floor material to see if there are other features ( e . g ., homeownership ) that could be leveraged to help implement the replacement of dirt floors with cement-like floors . Cement-like floors not only reduce the risk of Chagas disease by eliminating a hiding place for T . dimidiata nymphs ( Zeledón , Zúñiga , & Swartzwelder , 1969 ), they also help prevent infection with hookworm ( Hotez , 2008 ), which is the only neglected tropical disease with a higher DALY than Chagas in central Latin America ( Murray et al ., 2012 ). The cost of replacing dirt floors with cement-like floors is ~$ 170 ( Méndez , 2008 ). Guatemala has the highest prevalence of hookworm in Central
136