Phrase mining is a fundamental text mining task that aims to identify quality phrases from context. Nevertheless, the scarcity of extensive gold labels datasets, demanding substantial annotation efforts from experts, renders this task exceptionally challenging. Furthermore, the emerging, infrequent, and domain-specific nature of quality phrases presents further challenges in dealing with this task. In this paper, we propose LMPhrase, a novel unsupervised context-aware quality phrase mining framework built upon large pre-trained language models (LMs). Specifically, we first mine quality phrases as silver labels by employing a parameter-free probing technique called Perturbed Masking on the pre-trained language model BERT (coined as Annotator). In contrast to typical statistic-based or distantly-supervised methods, our silver labels, derived from large pre-trained language models, take into account rich contextual information contained in the LMs. As a result, they bring distinct advantages in preserving informativeness, concordance, and completeness of quality phrases. Secondly, training a discriminative span prediction model heavily relies on massive annotated data and is likely to face the risk of overfitting silver labels. Alternatively, we formalize phrase tagging task as the sequence generation problem by directly fine-tuning on the Sequence-to-Sequence pre-trained language model BART with silver labels (coined as Generator). Finally, we merge the quality phrases from both the Annotator and Generator as the final predictions, considering their complementary nature and distinct characteristics. Extensive experiments show that our LMPhrase consistently outperforms all the existing competitors across two different granularity phrase mining tasks, where each task is tested on two different domain datasets.

Extracting frequent words from a collection of texts is commonly performed in many subjects. However, as useful as it is to obtain a collection of commonly occurring words from texts, there is a need for more specific information to be obtained from texts in the form of most commonly occurring phrases. Despite this need, extracting frequent phrases is not commonly done due to inherent complications, the most significant being double-counting. Double-counting occurs when words or phrases are counted when they appear inside longer phrases that themselves are also counted, resulting in a selection of mostly meaningless phrases that are frequent only because they occur inside frequent super phrases. Several papers have been written on phrase mining that describe solutions to this issue; however, they either require a list of so-called quality phrases to be available to the extracting process, or they require human interaction to identify those quality phrases during the process. We present here a method that eliminates double-counting via a unique rectification process that does not require lists of quality phrases. In the context of a set of texts, we define a principal phrase as a phrase that does not cross punctuation marks, does not start with a stop word, with the exception of the stop words "not" and "no", does not end with a stop word, is frequent within those texts without being double counted, and is meaningful to the user. Our method identifies such principal phrases independently without human input, and enables their extraction from any texts within a reasonable amount of time.

Phrase mining is a key research problem for semantic analysis and text-based information retrieval. The existing approaches based on NLP, frequency, and statistics cannot extract high quality phrases and the processing is also time consuming, which are not suitable for dynamic on-line applications. In this paper, we propose an efficient high-quality phrase mining approach (EQPM). To the best of our knowledge, our work is the first effort that considers both intra-cohesion and inter-isolation in mining phrases, which is able to guarantee appropriateness. We also propose a strategy to eliminate order sensitiveness, and ensure the completeness of phrases. We further design efficient algorithms to make the proposed model and strategy feasible. The empirical evaluations on four real data sets demonstrate that our approach achieved a considerable quality improvement and the processing time was 2.3X - 29X faster than the state-of-the-art works.