This zip includes:

• Source code of our 2 streaming algorithms, greedy and SGr.
• Facebook dataset (in "data" folder) for testing the algorithms. Due to file size restriction in Github, please find the Sensor dataset in Intel Lab Data (http://db.csail.mit.edu/labdata/labdata.html).

Since estimating F in influence maximization is very time consuming, our code uses OpenMP for parallelization (https://en.wikipedia.org/wiki/OpenMP).

To build our code, run: g++ -std=c++11 *.cpp -o ksub -DIL_STD -fopenmp -g

After building, to run our code, run: ./ksub -f -V -t <type of experiment, 0: influence maximization, 1: sensor placement> -k -B -M -e -n <value of eta - denoise step for RStream> -g -a <algorithm, 0: greedy, 1: DStream, 2: RStream, 3: SGr. Please use SSA source code for testing IM algorithm> -p <number of threads (OpenMP) to running algorithms>

We conducted experiments on a Linux machine with 2.3Ghz Xeon 18 core processor and 256Gb of RAM. With 70 threads, DStream usually terminates after 20 minutes, RStream is 2 hours.