Tracking Idea Flows between Social Groups
Authors: Yangxin Zhong, Shixia Liu, Xiting Wang, Jiannan Xiao, Yangqiu Song
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on both synthetic and real datasets show that our method is more effective than methods based on traditional clustering techniques and achieves better accuracy. A case study was conducted to demonstrate the usefulness of our method in helping users understand the flow of ideas between different user groups on social media. |
| Researcher Affiliation | Academia | 1School of Software, Tsinghua University, Beijing, P.R. China 2Lane Department of Computer Science and Electrical Engineering, West Virginia University, United States |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks (clearly labeled algorithm sections or code-like formatted procedures). |
| Open Source Code | No | The paper does not provide concrete access to source code for the methodology described. There is no mention of a specific repository link, explicit code release statement, or code in supplementary materials. |
| Open Datasets | Yes | The experiment was conducted on ten real-world time series benchmark datasets in the UCR archive (Chen et al. 2015), in which the class label of each time series is known. |
| Dataset Splits | No | The paper mentions 'grid search on τmax ({1, 2, ..., 10})' and running 'each experiment 100 times with different random seeds' which implies internal validation, but it does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning into training, validation, and test sets. |
| Hardware Specification | Yes | All the experiments were conducted on a workstation with an Intel Xeon E52630 CPU (2.4 GHz) and 64GB of Memory. |
| Software Dependencies | No | The paper mentions algorithms and methods (e.g., K-means, BCC, DTW, greedy PARAFAC) but does not provide specific ancillary software details like library or solver names with version numbers needed to replicate the experiment. |
| Experiment Setup | Yes | In this experiment, synthetic datasets with five different noise levels were generated (noise level L {0, 0.2, ..., 0.8}). For each L, we generated 50 synthetic datasets and averaged the results of these datasets. For each dataset, we first generated ideas and the lead-lag relationships between ideas (Fig. 1(b)), which contain ck and Δtk. Then we generated the augmented bipartite word graph (Fig. 1(a)) according to noise level L. L denotes the probability that ck = 0 when ck = 1. The larger the L, the more difficult it is to generate accurate idea flows by using the augmented bipartite word graph. In the synthetic datasets, the number of ideas in each user group varied from 2 to 6, the number of words in each idea varied from 10 to 30, the lead-lag time varied from -6 to 6, and the number of time points in the lead or lag periods varied from 20 to 40, all while T was set to 200. We applied our algorithm with different tensors (X(1), X(2), and X(3)) to the datasets and reported the accuracy and efficiency. |