Skip to main content

Table 1 Significant associations with RF, BRT, LASSO and UFMLR

From: Detection of independent associations in a large epidemiologic dataset: a comparison of random forests, boosted regression trees, conventional and penalized logistic regression for identifying independent factors associated with H1N1pdm influenza infections

 

RF

BRT

LASSO-max

LASSO-se

UFMLR05

UFMLR05 backward

UFMLR20

UFMLR20 backward

Covariates

        

Pre-seasonal HAI titer (log)

0.004

0.003

0.001

0.001

0.001

0.002

0.002

0.001

History of asthmab

0.009

0.012

0.001

0.001

0.009

0.019

 

0.004

Professional activity involves contact with ill peopleb

0.009

0.023

0.008

   

0.001

0.001

Age (years)

 

0.004

0.001

0.001

 

0.001

  

Daily frequency of hand washing (with soap or hand sanitizer) ≥ 5b

0.014

0.023

0.019

    

0.008

Always or often covers mouth while coughing or sneezingb

0.033

0.023

0.003

0.001

    

“Craftsman, shopkeeper, chief executive officer” (socio-professional group)b

 

0.036

0.006

   

0.047

 

History of chemotherapyb

0.048

 

0.021

   

0.027

 

Average living room temperature (°C)

  

0.001

0.001

 

0.031

0.047

0.002

Presence of a dishwasher in the kitchenb

  

0.002

0.001

0.048

0.007

 

0.007

Sex = maleb

0.048

       

Professional activity is primarily outdoorsb

0.045

       

Age < 15 yearsb

0.016

       

Any respiratory diseaseb

0.030

       

Number of children (<15 years) in the household (n)

0.004

       

Body mass index (kg/m2)

 

0.032

      

Proportion of inhabitants > 15 years without a diploma (in IRIS zonei)

 

0.024

      

Proportion of habitations rented by inhabitants (in IRIS zonei)

 

0.023

      

Proportion of habitations owned by inhabitants (in IRIS zonei)

 

0.049

      

Habitation = houseb

 

0.016

      

Duration of contacts with subjects aged between 60 and 99 years (log (min))

  

0.006

   

0.023

0.019

Number of subjects in the household (n)

  

0.002

0.002

    

Kitchen surface area per subject (m2)

  

0.005

0.002

    

Cardiac arrhythmiab

  

0.021

     

History of radiotherapyb

  

0.016

     

Daily consumption of green tea (n)

  

0.040

     

Number of birds inside habitation (n)

  

0.030

     

Number of rooms per subject in habitation (n)

  

0.009

     

Number of children in the bedroom (n)

  

0.010

     

Bedroom windows face: gardenb

  

0.010

     

Duration of contacts at home (log (min))

  

0.029

     

Longitude of the habitation (degrees)

  

0.027

     

Latitude of the habitation (degrees)

  

0.028

     

Proportion of “farmer, primary sector” (socio-professional group) (in IRIS zonei)

  

0.004

    

0.005

Kitchen filtration of areab

      

0.021

0.034

Tiles flooring in the kitchenb

      

0.015

0.029

Agricultural land near habitationb

      

0.048

 
  1. bbinary covariate. iIRIS zones are statistical block groups of about 2000 inhabitants defined by the French Institut National de la Statistique et des Etudes Economiques (INSEE).