clessn/datagotchization

datagotchization's Issues

generate_ordinal_model()

## base model with all observations
	base_model <- model <- ordinal(vote ~ terms_config, data = data)
	## split data 5 times
	split_index <- caret::createDataPartition(data$vote, p = 0.75, list = FALSE, times = 5)
	for (i in 1:5){
		## each iteration, generate model on train
		train <- data[split_index[i, ], ]
		test <- data[-split_index[i, ], ]
		modeli <- ordinal(vote ~ terms_config, data = train)
		## add test data to the model
		modeli[["testdata"]] <- test
		## create a list with the 5 models
		if (i == 1){
			models <- list()
			models[[paste0("model", i)]] <- modeli
		} else {
    			models[[paste0("model", i)]] <- modeli
  		}
	## add it to the model object
	model[["cross_validation"]] <- models
	return(model)

generate_random_app_input()

generate_terms_configs()

1. Filter possible interactions term

2. Generate random list of {n_configs} terms_configs
   Random number of variables
   Random variables variation
   Random interactions terms
   Random zip code variables

   If we already have the config: retry
   If all the configs have been found: stop

3. Return list