Influence Maximization Documentation

• Author: Justin Wong justinryanwong@berkeley.edu

soc-digg dataset

After downloading and unzipping the data set, place the soc-digg folder into the Data folder. The path should include

# Data Directory Structure

The `Data` directory should have the following structure:

- `Data/Soc-digg/`
  - `soc-digg.mtx`
  - `readme.html`

Make sure to place the `soc-digg` folder into the `Data` directory after downloading and unzipping the dataset.

Running the Preprocess for IMM script

To run,

cd models
python3 preprocess_for_imm.py --fn soc-digg --source-network Data/soc-digg/soc-digg.mtx

This will generate the following files:

Generated Files

After running the preprocess_for_imm.py script, the following files will be generated:

• Data/soc-digg/output/wc_soc-digg_network.csv: CSV file containing the weighted cascade network data.
• Data/soc-digg/output/soc-digg_incr_dic.json: JSON file mapping incremental node ids for IMM.
• Data/soc-digg/output/wc_soc-digg_attribute.txt: Text file containing attributes for the network.

Additionally, a log file will be created in the logs directory with the format preprocess_for_imm-<timestamp>.log.

Overall, the output will look like

|── models/Data/
|        |── soc-digg/
|            |── output/
|                |── soc-digg_incr_dic.json
|                |── wc_soc-digg_attribute.txt
|                |── wc_soc-digg_network.csv
|            |── soc-digg.mtx
|        |── readme.html
|── models/logs/`
|        |── preprocess_for_imm-<timestamp>.log