FARMS: A New Algorithm for Variable Selection

Figure 2

Illustration of FARMS input and output. (a) FARMS function requires data as data frame containing all the variables (response and explanatory variables). (b) Calling the FARMS function within requires the indication of at least the response variable and the explanatory variables. In this illustration, which follows the explanation of Figure 1, we also indicate the number of variables to add in each iteration, the number of variables to compose the starting model, the columns containing the forced-in variables, and the name of the output file. (c) By default, FARMS function returns an object containing information for each iteration (two iterations in this illustration) and the dataset as processed by the algorithm. (d) Optionally, FARMS function can produce a text file containing the same information as the object output, adding a text line for each iteration, helping also to monitor the algorithm execution and to track the evolution of models until the final model is obtained.

(a) Example of dataset input format

(b) Example of FARMS function call

(c) Example of output from FARMS function within environment

(d) Example of TXT file output from FARMS function