On Estimating Recommendation Evaluation Metrics under Sampling

Authors: Ruoming Jin, Dong Li, Benjamin Mudrak, Jing Gao, Zhi Liu4147-4154

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we report the experimental evaluation on estimating the top-K metrics based on sampling, as well as the learning of empirical rank distribution P(R). Specifically, we aim to answer the following questions: (Question 1) How do the new estimators based on the learned empirical distribution perform against the CLS and BV approach proposed in (Krichene and Rendle 2020) on estimating the top-K metrics based on sampling? ... We use four of the most commonly used datasets for recommendation studies in our study...
Researcher Affiliation Collaboration Ruoming Jin,1 Dong Li, 1 Benjamin Mudrak,1 Jing Gao, 2 Zhi Liu2 1 Kent State University 2 i Lambda {rjin1,dli12,bmudrak1}@kent.edu {jgao,zliu}@ilambda.com
Pseudocode No The paper describes methods using mathematical equations and textual explanations but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not contain an explicit statement about the release of its source code for the methodology, nor does it provide a link to a code repository.
Open Datasets Yes We use four of the most commonly used datasets for recommendation studies in our study... (ml-1m dataset (Harper and Konstan 2015))... Table 3: Dataset: citeulike with sample size =99.
Dataset Splits No The paper mentions 'testing dataset' and 'sampling ranked results', and refers to sample sizes related to the evaluation process (e.g., 'sample size =99'), but it does not provide explicit details about train, validation, and test dataset splits (e.g., percentages or counts).
Hardware Specification No The paper does not provide specific details about the hardware (e.g., CPU/GPU models, memory) used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., programming languages, libraries, or frameworks with their versions) that would be needed to replicate the experiments.
Experiment Setup Yes The estimators include CLS, BV (with the tradeoff parameters \u03b3 = 0.1 and \u03b3 = 0.01), MLE (Maximal Likelihood Estimation), WMLE (Weighted Maximal Likelihood Estimation where the weighted function is MNDCG with C = 10), MES (Maximal Entropy with Squared distribution distance, where \u03b7 = 0.001).