Identifying Key Observers to Find Popular Information in Advance

Authors: Takuya Konishi, Tomoharu Iwata, Kohei Hayashi, Ken-ichi Kawarabayashi

IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In experiments, we test our approach using real social bookmark datasets. The results demonstrate that our approach can find popular items in advance more effectively than baseline methods.
Researcher Affiliation Collaboration National Institute of Informatics JST, ERATO, Kawarabayashi Large Graph Project NTT Communication Science Laboratories
Pseudocode Yes Algorithm 1 Ada-RDA (E, λ, , )
Open Source Code No The paper does not provide any explicit statements about releasing source code or links to a code repository for the described methodology.
Open Datasets Yes We used Delicious datasets [Wetzker et al., 2008], which comprise records of events where Delicious users bookmarked (adopted) web pages with time stamps.
Dataset Splits Yes Items were split into ten subsets, with 90 percent of the items used as training data and the other 10 percent as test data. ... We repeated the above procedure ten times while changing the training and test data (i.e. 10-fold cross validation) and took the average of the AUCs.
Hardware Specification Yes We used one server that has 16 processors and allows for computing 32 threads at a time by hyper-threading.
Software Dependencies No While the proposed methods were implemented by Java, baselines were done by Python. The paper does not provide specific version numbers for any software dependencies.
Experiment Setup Yes We set both and to 1.0, and C to 20. We also set λ to the set of 500 points in [0.00001, 0.01]. For the augmented method, we needed to select the pair of parameters (k, ). We performed v-fold cross validation on k ∈ K = {0.5, 1.0, 2.0, 5.0, 10.0, 50.0}, σ ∈ {1.0, 5.0, 9.5, 15.0}, and v = 5 using training data.