DBLP GraphDatabase

Loading

1. Generate CSV file of the DBLP XML dump file using ThomHurks tools (https://github.com/ThomHurks/dblp-to-csv) and store it in the "dblp_dump/" directory in this project.
   python XMLToCSV.py dblp.xml dblp.dtd output.csv
2. Run a2_data_preparation_NugrohoSalmaan.py
3. Copy the output in the "out/" directory to neo4j import directory, by default the directory path is /var/lib/neo4j/import/ Run a2_loading_NugrohoSalmaan.py
4. Run a2_loading_NugrohoSalmaan.py

Evolving

1. Run a3_prepare_evolving.py
2. Copy the the generated file (out/authors_with_affiliation.csv) to neo4j import directory
3. Run a3_evolving_NugrohoSalmaan.py to start the evolving process

Running Queries

Run b_queries_NugrohoSalmaan.py using -q parameter specifying query number. For example to run query 1 (the h-index query), run the following command:

python b_queries_NugrohoSalmaan.py -q 1

The fourth query require additional year parameter that can be specified using -y option. For example:

python b_queries_NugrohoSalmaan.py -q 4 -y 1999

Running Graph Algorithm

Run c_graph_algorithm_NugrohoSalmaan using -a parameter specifying algorithm number. To run our first algorithm (article rank) use -a 1, and to run our second algorithm (louvain community detection) use -a 2. For example:

python c_graph_algorithm_NugrohoSalmaan.py -a 1

Running Recommender

Run d_recommender_NugrohoSalmaan.py with the following parameters:

• -t (in a fraction) homogenity threshold for a conference to be considered as community
• -m minimum paper appears in top 100 to be considered as guru
• -k string representation of keywords, e.g:'["database", "sql"]'

For example:

python d_recommender_NugrohoSalmaan.py -t 0.5 -m 1 -k '["database", "sql"]' 

In order to get meaningful result, we can use the following keyword list as keywords argument:

'["database","index","indexing","querying","data","sql","olap","b+-trees","dbms","dbmss","views","databases","queries","postgres","picodbms","xml","tables","oodbms","relational"]'