On Multiset Selection With Size Constraints
Authors: Chao Qian, Yibo Zhang, Ke Tang, Xin Yao
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on budget allocation as well as a more complex application, namely, generalized influence maximization, exhibit the superior performance of the proposed approach. |
| Researcher Affiliation | Collaboration | 1Anhui Province Key Lab of Big Data Analysis and Application, School of Computer Science and Technology, University of Science and Technology of China, Hefei 230027, China 2Shenzhen Key Lab of Computational Intelligence, Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China |
| Pseudocode | Yes | Algorithm 1 POMS Algorithm |
| Open Source Code | No | The paper does not provide an explicit statement or link for the open-sourcing of their code. |
| Open Datasets | Yes | We use one real-world data set Yahoo! Search Marketing Advertiser Bidding Data1, which is a bipartite graph representing which customers are interested in which keywords . It contains n = 1,000 source nodes (i.e., keywords), 10,475 target nodes (i.e., customers) and 52,567 edges. For each influence probability p(j) i of each source node vi, we use a randomly generated value between 0 and 0.5. The capacities ci are set to 5 (thus cmax = 5). The budget k is set from 10 to 100. For the case with a competitor, the probability q(j) i is set to 0.2 p(j) i , and the budget of the competitor is set to 100, which is allocated to the top 100 highest degree source nodes in advance. We use two real-world data sets: ego-Facebook2 (4,039 nodes, 88,234 edges) and Weibo (5,000 nodes, 65,148 edges). |
| Dataset Splits | No | The paper does not specify training, validation, and test splits. It uses different datasets for experiments but doesn't detail how they were partitioned for model training and evaluation cycles in terms of distinct sets for training, validation, and testing. |
| Hardware Specification | No | The paper does not specify the hardware used for running the experiments (e.g., CPU, GPU models). |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers. |
| Experiment Setup | Yes | The number T of iterations of POMS is set to 2ecmaxk2n as suggested by Theorem 1. As POMS is a randomized algorithm, we repeat the run 10 times independently and report the average results. For each influence probability p(j) i of each source node vi, we use a randomly generated value between 0 and 0.5. The capacities ci are set to 5 (thus cmax = 5). The budget k is set from 10 to 100. For the case with a competitor, the probability q(j) i is set to 0.2 p(j) i , and the budget of the competitor is set to 100, which is allocated to the top 100 highest degree source nodes in advance. On each network, the probability of one edge from vi to vj is estimated by weight(vi,vj) indegree(vj). The capacities ci are set to 5. The budget k is set from 5 to 10. To estimate the expected number of active nodes, we simulate the diffusion process 30 times independently and use the average as an estimation. But for the final output solutions of the algorithms, we average over 10,000 times for more accurate estimation. |