select.people: select.people

In HumanExposure/SHEDSDevel: The SHEDS-HT model for estimating human exposure to chemicals.

Description

Assigns demographic and physiological variables to each theoretical person to be modeled.

Usage

select.people(n, pop, py, act.p, diet.p, act.d, diet.d, specs)

Arguments

n	Number of persons.
pop	The population input; output of read.pop.file function. Contains counts by gender and each year of age from the 2000 U.S. census. When a large age range is modeled, this ensures that SHEDS chooses age and gender with the correct overall probability. #'
py	Regression parameters on the three physiological variables of interest (weight, height, and body mass index) for various age and gender groups. Output of the read.phys.file function.
act.p	Activity diary pools; output of the act.diary.pools function. Each element in this input is a list of acceptable activity diary numbers for each year of age, for each gender, weekend, and season combination.
diet.p	Dietary diary pools; output of the diet.diary.pools function. Each element in this input is a list of acceptable activity diary numbers for each year of age, for each gender, weekend, and season combination.
act.d	Activity diaries, which indicate the amount of time and level of metabolic activity in various 'micros'. Each line of data represents one person-day (24 hours). Output of the read.act.diaries function.
diet.d	Daily diaries of dietary consumption by food group; output of the read.diet.diaries function. Each line represents one person-day, with demographic variables followed by amounts (in grams/day) for a list of food types indicated by a short abbreviation on the header line.
specs	Output of the read.run.file function, which can be modified by the update.specs function before input into select.people.

Details

This is the first real step in the modeling process. It first fills an array q with uniform random numbers, with ten columns because there are 10 random variables defined by this function. There number of rows correspond to the number of persons, capped at the set.size specified in the Run input file (typically 5000). Gender is selected from a discrete (binomial) distribution where the counts of males and females in the study age range determines the gender probabilities. Age is tabulated next, separately for each gender. The counts by year of age are chosen for the appropriate gender and used as selection weights. Season is assigned randomly (equal weights) using those specified in the Run input file. Weekend is set to one or zero, with a chance of 2/7 for the former. The next block of code assigns physiological variables. Weight is lognormal in SHEDS, so a normal is sampled first and then exp() is applied. This means that the weight parameters refer to the properties of log(weight), which were fit by linear regression. The basal metabolic rate (bmr), is calculated by regression. A minimum bmr is set to prevent extreme cases from becoming zero or negative. The alveolar breathing ventilation rate corresponding to bmr is also calculated. The SHEDS logic sets activities in each micro to be a multiple of these rates, with outdoor rates higher than indoor, and indoor rates higher than sleep rates. This calculated activities affect the inhaled dose. The skin surface area is calculated using regressions based on height and weight for 3 age ranges. The next step is to assign diaries. Here, a FOR loop over n persons (n rows) is used to assign appropriate diet and activity diary pools to each person. An empirical distribution is created, consisting of the list of diary numbers for each pool. The final step is to retrieve the actual data from the chosen activity and diet diaries, and the result becomes pd.

Value

pd A dataframe of "person-demographics": assigned demographic and physiological parameters for each theoretical person modeled in SHEDS.HT

Author(s)

Kristin Isaacs, Graham Glen

See Also

run, read.pop.file, read.phys.file, act.diary.pools, diet.diary.pools, read.act.diaries, read.diet.diaries, update.specs

HumanExposure/SHEDSDevel documentation built on Oct. 30, 2019, 6:49 p.m.