predict2: Takes predictions from model object(s) and and combines it...

In bioticinteractions/predict2: Take model objects to make predictions, save model objects and predictions to disk

Description

Takes predictions from model object(s) and and combines it with actual values This function takes a vector of strings that represent the model object names

Usage

predict2(model, newdata, actual, pred_type, append_cols, write_model,
  write_pred, csv_name, dir, dir_data, dir_csv)

Arguments

model	A vector of containing strings of the model object names
newdata	A dataframe with input variables that predict2() will feed into the model
actual	Vector of expected (or 'actual') data from the test dataset. Will be merged to the predicted data for easy export and subsequent comparison
pred_type	The type of predicted value predict will return: regression will return values; classification options: response prob, vote. Default is 'response' [see predict.randomForest for more information]
append_cols	Append additional columns to the predictions .csv.
write_model	Logical. Whether to write the model fit objects to disk (as .data; one for each model). Default = FALSE
write_pred	Logical. Whether to write the predictions to disk (as a .csv). Default = FALSE
csv_name	String add to the fileneames of the data and prediction files (Default value = 'model'; model.data; model.csv)
dir	Path location where the .csv and .data files will be written. Default is current working directory: getwd()
dir_data	Specify location for data files if you want the data files to be in a separate file. Default is current working directory: getwd()
dir_csv	Specify location for csv files if you want the data files to be in a separate file. Default is current working directory: getwd() working directory (i.e. getwd())

Value

predict2() will return the results from predict() as a data.frame. each model object in model' will be one column in the dataframe.

Examples

library(predict2)
data(lm_data)
df_train = lm_data[1:25, ]
df_test = lm_data[26:50, ]
lm_int = lm(y ~ x, data = df_train)
lm_noint = lm(y ~ 0 + x, data = df_train)
list_lm = c('lm_int', 'lm_noint')

predicted = predict2(
  model = list_lm,
  newdata = df_test,
  actual = df_test$y,
  append_cols = df_test,
  pred_type = 'response',
  write_model = FALSE,
  write_pred = FALSE,
  csv_name = 'none'
)
head(predicted)

bioticinteractions/predict2 documentation built on May 28, 2019, 7:12 p.m.