Multi-Label Classification with Feature-Aware Non-Linear Label Space Transformation

Authors: Xin Li, Yuhong Guo

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on a number of multi-label classification datasets. The proposed approach demonstrates good performance, comparing to a number of stateof-the-art label dimension reduction methods.
Researcher Affiliation Academia Xin Li and Yuhong Guo Department of Computer and Information Sciences Temple University, Philadelphia, PA 19122, USA {xinli, yuhong}@temple.edu
Pseudocode Yes Algorithm 1 A Unified Training Algorithm
Open Source Code No No statement providing access to open-source code was found.
Open Datasets Yes We used five real world multi-label datasets for image and text categorization tasks in our experiments, including Corel5K, ESPGame, Iaprtc12, Enron, and Delicious.
Dataset Splits Yes We conducted experiments using 10-fold cross validation on four datasets, except the large scale dataset Delicious, on which we conducted experiments using 5-fold cross validation. In each cross validation iteration, we performed parameter selection for all the comparison methods by using 80% of the training set for training and the remaining 20% for performance evaluation.
Hardware Specification Yes To compare the empirical computational complexity of the comparison methods, we reported in Table 2 the training time and testing time of each method for a single run with θ=0.3 on a 64-bit PC with 4 processors (3.4 GHz) and 16 GB memory.
Software Dependencies No No specific software dependencies with version numbers were mentioned.
Experiment Setup Yes For the proposed COMB method, there are two parameters µ and γ to be tuned for the decoding process. We selected the µ value from the set [0.001, 0.005, 0.01, 0.05, 0.1], and selected the γ value from the set [0, 0.2, 0.4, 0.6, 0.8, 1].