Class Prior Estimation with Biased Positives and Unlabeled Examples

Authors: Shantanu Jain, Justin Delano, Himanshu Sharma, Predrag Radivojac4255-4263

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our empirical investigation suggests feasibility of the correction strategy and overall good performance. Experiments and Results summarize our empirical investigation, summarizing the datasets, experimental protocols and results.
Researcher Affiliation Academia Shantanu Jain, Justin D. Delano, Himanshu Sharma, Predrag Radivojac Khoury College of Computer Sciences Northeastern University, Boston, MA, U.S.A.
Pseudocode Yes Algorithm 1 Algorithm for class prior estimation with biased positives and unlabeled examples. // max K specifies the maximum number of clusters. Require: M, C, max K Ensure: α // Partition the biased positive set by k-means clustering. // The number of clusters is picked to be the one giving // a clustering with the maximum Silhouette coefficient, // up to a maximum of max K. c Part[i] stores the // positives in the ith cluster. c Part k Means Silhouette(C, max K)
Open Source Code No The paper does not provide explicit statements or links for the open-sourcing of the described methodology's code.
Open Datasets Yes Our experiments were carried out on twelve real-life datasets from the UCI Machine Learning Repository (Lichman 2013).
Dataset Splits No The paper describes the generation of biased and unbiased positive-unlabeled datasets but does not explicitly provide training, validation, and test splits with percentages or counts.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or cloud instances) used for running the experiments.
Software Dependencies No The paper mentions algorithms and libraries (e.g., 'k-means algorithm', 'Alpha Max', 'Elkan-Noto algorithm') but does not provide specific version numbers for software dependencies.
Experiment Setup Yes Corrected is an exact implementation of Algorithm 1 with max K intialized to 5. To generate biased positive examples and unlabeled data, the positive examples were clustered using k-means, where the number of clusters, K, was determined based on the Silhouette coefficient.