Fake Review Detection
A research area focused on identifying and filtering fraudulent or deceptive reviews on online platforms using machine learning, natural language processing, and related techniques. It encompasses multi-dimensional signal fusion methods spanning textual analysis, behavioral features, and graph-based structures.
Timeline (last 90 days)
A survey on fake review detection covers 211 research papers and systematically reviews the evolution from traditional machine learning to the large language model era.
The evidence sources summarized in the survey include review text, sentiment signals, rating behavior, temporal metadata, user–product relationship graphs, multimodal content, external knowledge, and LLM-generated signals.
Due to differing label definitions, a model that achieves high scores on one dataset may see substantially degraded performance on another, making cross-paper numbers not necessarily comparable.
The survey identifies six open challenges: adversarial generation, cross-domain transfer, uncertainty-aware fusion, missing-source robustness, interpretability, and trustworthy evaluation.
Effective fake review governance requires integrating multi-dimensional evidence from behavioral, structural, and temporal dimensions, and the era of relying solely on text classification models is passing.
All Facts (5)
A survey on fake review detection covers 211 research papers and systematically reviews the evolution from traditional machine learning to the large language model era.
50%UnverifiedThe evidence sources summarized in the survey include review text, sentiment signals, rating behavior, temporal metadata, user–product relationship graphs, multimodal content, external knowledge, and LLM-generated signals.
50%UnverifiedDue to differing label definitions, a model that achieves high scores on one dataset may see substantially degraded performance on another, making cross-paper numbers not necessarily comparable.
50%UnverifiedThe survey identifies six open challenges: adversarial generation, cross-domain transfer, uncertainty-aware fusion, missing-source robustness, interpretability, and trustworthy evaluation.
50%UnverifiedEffective fake review governance requires integrating multi-dimensional evidence from behavioral, structural, and temporal dimensions, and the era of relying solely on text classification models is passing.
50%