R语言机器学习代写 Applied regression Assignment 1

The University of Hong Kong School of Public Health

Master of Public Health Advanced Statistical Methods I

Applied regression Assignment 1

(submit to Moodle by 11:55pm, Mar 22, 2018)

March 12, 2018

Association of invasive meningococcal disease and environmental and virologic exposure

Weekly counts of invasive meningococcal disease (IMD) reported cases from country A in 1997-2000 were collected and analyzed, to study associations between IMD and environmental and virologic exposure. The file ‘IMD.csv’ stored the following variables which were found to be significant factors from the literature in other places:

 flu: influenza isolation rate from laboratory  maxtemp: weekly maximum temperature  IMD: weekly number of IMD cases
 pop: population size

a) Fit a poisson regression model to predict weekly IMD cases using flu and maxtemp, accounting for the increasing population size in 1997-2000. Summarize your results in a table. [4 marks]

  1. b)  Assess the goodness of fit of the model. [1 marks]
  2. c)  Quote the mean and variance of the weekly IMD cases. Is there any

evidence of overdispersion? [1 marks]

d) Fit a negative binomial regression model to predict weekly IMD cases and summarize your results in a table. [4 marks]

  1. e)  Assess the goodness of fit of the model. [1 marks]
  2. f)  Calculate the AIC for the poisson and negative binomial regression

models and select the best model based on AIC and goodness of fit. [1 marks]

g) Assess if the linear effects for the variables flu and maxtemp are adequate. [2 marks]

h) Assess if there is any collinearity problem in your final model? [2 marks]

i) Assess if there is any unexplained serial correlation after fitting your final model. [2 marks]

  1. j)  Draw an overall conclusion on the findings. [3 marks]
  2. k)  Based on the final model, predict the number of IMD cases in a week, if

the influenza isolation rate is 0.3, with a maximum temperature of 5oC and a population of 110 million. [2 marks]

l) Based on the final model, which of the conditions below will have a higher risk of IMD? [2 marks]

1. Influenzaisolationrate=0.3,maximumtemperature=5oC 2. Influenzaisolationrate=0.4,maximumtemperature=0oC