Papers › The Document Vectors Using Cosine Similarity Revisited

The Document Vectors Using Cosine Similarity Revisited

26 May 2022insights (ACL) 2022 5arXiv:2205.13357archive 2025-07-28

Zhang Bingyu, Nikolay Arefyev

The current state-of-the-art test accuracy (97.42%) on the IMDB movie reviews dataset was reported by \citet{thongtan-phienthrakul-2019-sentiment} and achieved by the logistic regression classifier trained on the Document Vectors using Cosine Similarity (DV-ngrams-cosine) proposed in their paper and the Bag-of-N-grams (BON) vectors scaled by Naive Bayesian weights. While large pre-trained Transformer-based models have shown SOTA results across many datasets and tasks, the aforementioned model has not been surpassed by them, despite being much simpler and pre-trained on the IMDB dataset only. In this paper, we describe an error in the evaluation procedure of this model, which was found when we were trying to analyze its excellent performance on the IMDB dataset. We further show that the previously reported test accuracy of 97.42% is invalid and should be corrected to 93.68%. We also analyze the model performance with different amounts of training data (subsets of the IMDB dataset) and compare it to the Transformer-based RoBERTa model. The results show that while RoBERTa has a clear advantage for larger training sets, the DV-ngrams-cosine performs better than RoBERTa when the labelled training set is very small (10 or 20 documents). Finally, we introduce a sub-sampling scheme based on Naive Bayesian weights for the training process of the DV-ngrams-cosine, which leads to faster training and better quality.

PaperPDFConference PDFCode

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

bgzh/dv_cosine_revisited officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Sentiment Analysis

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Sentiment Analysis IMDb DV-ngrams-cosine with NB sub-sampling + RoBERTa.base Accuracy 95.94 #7 of 49 Archive leaderboard report
Sentiment Analysis IMDb DV-ngrams-cosine + RoBERTa.base Accuracy 95.92 #8 of 49 Archive leaderboard report
Sentiment Analysis IMDb RoBERTa.base Accuracy 95.79 #12 of 49 Archive leaderboard report
Sentiment Analysis IMDb DV-ngrams-cosine + NB-weighted BON (re-evaluated) Accuracy 93.68 #25 of 49 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdamAttentionAttention DropoutBERTDense ConnectionsDropoutLayer NormalizationLinear LayerLinear Warmup With Linear DecayLogistic RegressionMulti-Head AttentionResidual ConnectionRoBERTaSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections