Personalized Sentiment Classification Based on Latent Individuality of Microblog Users

Authors: Kaisong Song, Shi Feng, Wei Gao, Daling Wang, Ge Yu, Kam-Fai Wong

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

Reproducibility Variable Result LLM Response
Research Type Experimental 5 Experimental Evaluation and Results on real-world microblog datasets confirm that our method outperforms stateof-the-art baseline algorithms with large margins.
Researcher Affiliation Academia 1Northeastern University, Shenyang, China 2Qatar Computing Research Institute, Qatar Foundation, Doha, Qatar 3The Chinese University of Hong Kong, Hong Kong
Pseudocode No No structured pseudocode or algorithm blocks were found in the paper.
Open Source Code No We implemented our models based on the generic factorization tool SVDFeature7. (Footnote 7: svdfeature.apexlab.org/wiki/Main_Page) - This refers to a tool used, not the authors' own code release.
Open Datasets No We crawled the microblog posts of 281 Sina Weibo users and 674 Twitter users using Weibo API2 and Twitter API3, respectively. We obtained 43,250 Weibo posts and 48,563 tweets, each containing one positive or negative emoticon. The paper does not provide a link or statement that these crawled datasets are publicly available. It only mentions the source of the raw data (APIs) and the lexicons (which are publicly available, but not the processed dataset itself).
Dataset Splits Yes We used 10-fold cross validation for evaluation, where 8 folds were for training, 1 for development and 1 for test.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No For both datasets, Stanford POS tagger and neural network dependency parser were employed for POS tagging and dependency parsing, respectively (see Section 4.1). The Chinese sentimental words ontology bank5 and NRC s Emo Lex and Max Diff Twitter Sentiment Lexicon6 were used as sentiment lexicons for Weibo posts and tweets, respectively. We implemented our models based on the generic factorization tool SVDFeature7. While software and tools are mentioned, specific version numbers for these software dependencies are not provided.
Experiment Setup Yes We optimized the f parameter via validation on the development set by performing a grid search on all values of 10 x with x {1, 2, ..., 10}. Basically the performance was not sensitive with respect to f, and we fixed f = 60 which is slightly better than other choices. We tuned λ using the development data and fixed it as 1.0e-4. Since λU and λT are calculated by δ2 /δ2 U and δ2 /δ2 T (see formula 4), we set the two ratios to fixed values as 1.0e-3 as we found that varying them just influenced the results slightly.