Exploiting Emotion on Reviews for Recommender Systems
Authors: Xuying Meng, Suhang Wang, Huan Liu, Yujun Zhang
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirical results on real-world datasets demonstrate the effectiveness of our proposed framework and further experiments are conducted to understand how emotion on reviews works for the proposed framework. In this section, we conduct experiments to demonstrate the effectiveness of our proposed framework MIRROR. |
| Researcher Affiliation | Academia | Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China University of Chinese Academy of Sciences, Beijing 100049, China Computer Science and Engineering, Arizona State University, Tempe, 85281, USA |
| Pseudocode | Yes | Algorithm 1 The Proposed Framework MIRROR |
| Open Source Code | No | The paper does not provide any explicit statement or link indicating that the source code for the methodology is openly available. |
| Open Datasets | Yes | We collected two real-world datasets Ciao and Epinions. For Ciao, users can rate products and reviews with scores from 1 to 5. For the evaluation purpose, we take the ratings on reviews as ground truth of emotion. In particular, we consider low review ratings from 1 to 3 as negative emotion, high review ratings 4 and 5 as positive emotion. For Epinions, the rating scope on reviews is from 1 to 6, where we regard ratings from 1 to 3 as negative emotion and 4 to 6 as positive emotion. |
| Dataset Splits | Yes | In datasets Ciao and Epinions, we randomly select x% of user ratings and corresponding emotion information related to the selected user ratings as the training set and the remaining 1 x% as the testing set. To investigate the capability of the proposed framework in handling the data sparsity problem, we vary x as {10, 20, 40} in this work. We then apply five fold cross validation for all the following experiments, and report the average MAE and RMSE. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware (e.g., CPU, GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper describes the methods used (e.g., Matrix factorization, t-test) but does not provide specific software names with version numbers or library dependencies for replication. |
| Experiment Setup | Yes | For MIRROR, we empirically set α = 0.01, β = 0.1, γ = 0.5, K = 10. More details about parameter selection for MIRROR will discussed in the following subsections. |