kojonketia/Using-hurdle-models-to-analyze-schistosomiasis-count-data

The count data and model codes for analysis for the paper "Using zero-inflated and hurdle models to analyze schistosomiasis count data of school children in the southern areas of Ghana" are presented here. The codes are a combination of Python and R codes. # Using-hurdle-models-to-analyze-schistosomiasis-count-data The excel file named "schistosomiasis_count_data.xlsx" contains the demographic features and the "Dataset.xlsx" is the dataset after extra columns where certain variables were grouped and it is what was used for analysis. ## Files The Rmd file contains the model codes used. Six distinct models under different scenarios were used. The python (jupyter) file contains codes used for slicing dataframes, descriptive statistics and fugure plotting ## Methods ### Rmd The standard Poisson, negative binomial (NB), zero-inflated Possion (ZIP), zero-inflated negative binomial (ZINB), hurdle Poisson (HP), hurdle negative binomial (HNB) models are the various models used. Their corresponding R syntaxes are showing the table that follows below | Model | Syntax | | --- | --- | | Poisson | glm("dependent variable ~ x1 + x2 + ... + xn", family = 'poisson', data = count_data) | | NB | glm("dependent variable ~ x1 + x2 + ... + xn", family = negative.binomial(1), data = count_data | | ZIP | zeroinfl(dependent variable ~ x1 + x2 + ... + xn, data = count_data, dist = "poisson") | | ZINB | zeroinfl(dependent variable ~ x1 + x2 + ... + xn, data = count_data, dist = "negbin") | | HP | hurdle(dependent variable ~ x1 + x2 + ... + xn, data = count_data, dist = "poisson") | | HNB | hurdle(dependent variable ~ x1 + x2 + ... + xn, data = count_data, dist = "negbin") | The results output of the each model is generated by "summary(model)" ``` # using the Possion model poisson 15), area (rural and urban) and educational level (preschool, primary and junior high) respectively. #### Model's AIC values In this section the AIC values of all scenarios used under different models were uploaded for visual comparison #### Observed and model's expected values The folder "model results" contains the each model's expected values. These values are then plotted against the observed values in 1) a bars against each other in log aixs and 2) …