README.md

SpecNFS dataset is a collection of 1198 NFS-RFC specifications annotated with their SpecIR representation.

The datasets in SpecNFS along with their brief description are as follows:

1. ./dataset/NER_dataset.csv

   This is the dataset containing the information about the spans and type of the entities in a sentence

   Column	Description
   Sentence	Sentence number
   Token	Token in the sentence
   Token_ID	Unique identifier of the span of entity in a sentence
   Token_no	Sequence number of a token in a sentence.
   Partition	The partition no. of the sentence which was used for the 5-fold cross validation reported in the paper

2. ./dataset/Link_dataset.csv

   This is the dataset containing dependency link information between entities in a sentence

   Column	Description
   Head_id	Entity span identifier of the head of the link
   Dep_id	Entity span identifier of the dependent of the link
   Link	Type of link between the head and the dependenct entity
   Sentence_no	Sentence number of the link
   Partition	The partition no. of the sentence which was used for the 5-fold cross validation reported in the paper

3. ./dataset/token_type2link_type.csv

   This is the dataset containing the valid link types between different pair of entity types

   Column	Description
   H_Token_TYPE	Head entity type
   D_Token_TYPE	Dependenct entity type
   Link_TYPE	Valid link types between head and dependenct entity types