Efficient Approximation Algorithms for the Diameter-Bounded Max-Coverage Group Steiner Tree Problem

It contains all the data used in our experiments, including 5 real graphs (MONDIAL, OpenCyc, LinkedMDB, YAGO, and DBpedia) and 3 synthetic graphs (LUBM-50K, LUBM-500K, and LUBM-5M).

Each real KG directory contains 8 files, including:

graph.txt: The first value is the number of vertices. Then each line 'u v' means there is an undirected edge between 'u' and 'v'.

Weightgraph.txt: The first value is the number of vertices. Then each line 'u v w' means there is an undirected edge between 'u' and 'v' weighted by 'w' which is computed by the Informativeness-based Weighting (IW) scheme.

nodeName.txt: Mapping from vertex ID to vertex name (i.e., entity URI).

query.txt: Each line is a keyword query containing a set of keyword names.

kwName.txt: Mapping from keyword ID to keyword name.

kwMap.txt: Mapping from keyword ID to vertex IDs. The first value of each line is keyword ID, and the rest are vertex IDs.

UWHBLL.txt: The HBLL index file which was built based on the Unit Weighting.

IWHBLL.txt: The HBLL index file which was built based on the Informativeness-based Weighting.

Each synthetic directory contains 6 files, including:

graph.txt: same as above.

Weightgraph.txt: same as above.

nodeName.txt: same as above.

queryList.txt: Each line contains a set (separated by ',') of sets of vertex IDs.

UWHBLL.txt: same as above.

IWHBLL.txt: same as above.

Apart from that, Dbpedia and LUBM-5M also contain a PLLlabel.txt file which was the supplementary file for the HBLL index.