Multi Column Description Format

The Multi Column Description Format (mcd) associates labels to the columns of an mcf file. Each line of an mcd file describes a column of an mcf file. Each line of an mcd file is made of four columns

• The first column contains an integer that indicates the column of the mcf that is being described
• The second column associates a label to the mcf column that is being described
• The third column indicates the type of the value found in the column. Three types are defined:
  • VOCAB indicates that the values are labels
  • INT indicates that the values are integers
  • EMB indicates that the values are real valued vectors (embeddings)
• The fourth column is used for embeddings, it is the name of a file containing the embeddings

Here is an example of an mcd file

1 | FORM | VOCAB | _ ---- | ---- | ---- | ---- | ---- 2 | POS | VOCAB | _ 3 | LEMMA | VOCAB |_ 4 | GOV | INT | _ 5 | LABEL | VOCAB | _ 6 | SENT_SEG | INT | _

An example of an mcf corresponding to this mcd cand be found here