Learning Algorithms for Keyphrase Extraction

By Peter D. Turney. Information retrieval (Boston), 2000

943 citations recorded by the source corpus.

Many academic journals ask their authors to provide a list of about five to fifteen keywords, to appear on the first page of each article. Since these key words are often phrases of two or more words, we prefer to call them keyphrases. There is a wide variety of tasks for which keyphrases are useful, as we discuss in this paper. We approach the problem of automatically extracting keyphrases from text as a supervised learning task. We treat a document as a set of phrases, which the learning algorithm must learn to classify as positive or negative examples of keyphrases. Our first set of experiments applies the C4.5 decision tree induction algorithm to this learning task. We evaluate the performance of nine different configurations of C4.5. The second set of experiments applies the GenEx algorithm to the task. We developed the GenEx algorithm specifically for automatically extracting keyphrases from text. The experimental results support the claim that a custom-designed algorithm (GenEx), incorporating specialized procedural domain knowledge, can generate better keyphrases than a general-purpose algorithm (C4.5). Subjective human evaluation of the keyphrases generated by GenEx suggests that about 80% of the keyphrases are acceptable to human readers. This level of performance should be satisfactory for a wide variety of applications.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/591394cfab5879934a59f6c44f03b6c4009042f0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

Citations in the corpus

The battle of the buzzwords: A comparative review of the circular economy and the sharing economy concepts (2021, methodology): “…future generations (Liu, 2012; Martins, 2016; 6 Keywords plus or Index keywords – search filters automated by scientific databases based on algorithms scanning e.g. an article’s references’ titles (Garfield, 1990; Gil-Leiva and Alonso-Arroyo, 2007; Turney, 2000) – were not included in the research.”
Automatic Keyphrase Extraction via Topic Decomposition (2010, methodology): “Automatic keyphrase extraction is widely used in information retrieval and digital library (Turney, 2000; Nguyen and Kan, 2007).”
Single Document Keyphrase Extraction Using Neighborhood Knowledge (2008, background): “GenEx (Turney, 2000) and Kea (Frank et al., 1999; Witten et al., 1999) are two typical systems, and the most important features for classifying a candidate phrase are the frequency and location of the phrase in the document.”
A Language Model Approach to Keyphrase Extraction (2003, background): “Some work has been done in extracting keyphrases from technical documents treating keyphrase extraction as a supervised learning problem (Frank et al., 1999; Turney, 2000).”
A New Approach to Improving Multilingual Summarization Using a Genetic Algorithm (2010, methodology): “, 1997), sentences (Luhn, 1958), keyphrases (Turney, 2000) or keywords (Litvak and Last, 2008).”
Evaluating the use of search engine development tools in IT education (2010, neutral): “A detailed description of the Extractor keyphrase extraction algorithm is provided in Turney (2000). The relatively low number of terms (30 or less) used by the Extractor to represent a document prevents the need to apply a dimensionality reduction process or handle long vectors, which could increas”
Interpretation and trust: designing model-driven visualizations for text analysis (2012, background): “Others note that single words account for only a small fraction of descriptive phrases used by people [48].”
Wikify!: linking documents to encyclopedic knowledge (2007, methodology): “Turney [28] evaluated a supervised system on a corpus of journal articles and re-”
Improved Automatic Keyword Extraction Given More Linguistic Knowledge (2003, methodology): “Turney (2000) presents results for a comparison be-tween an extraction model based on a genetic algo-rithm and an implementation of bagged C4.5 deci-216 sion trees for the task.”
Finding advertising keywords on web pages (2006, methodology): “One of the best known programs for keyword extraction is Turney’s GenEx system [22].”
This Email Could Save Your Life: Introducing the Task of Email Subject Line Generation (2019, background): “…and McKeown, 2004; Carenini et al., 2007; Zajic et al., 2008; Carenini et al., 2008; Ulrich et al., 2009), keyword extraction and action detection (Turney, 2000; Ben-nett and Carbonell, 2005; Dredze et al., 2008; Scerri et al., 2010; Loza et al., 2014; Lahiri et al., 2017; Lin et al., 2018), and…”
Deep multiple instance learning for image classification and auto-annotation (2015, methodology): “There have been abundant classical techniques to extract keywords from texts [40].”
Discovering key concepts in verbose queries (2008, methodology): “Instead, we adopt the supervised machine learning approach [29, 12, 15] for key query concept extraction, and use a diverse mix of features as inputs for the supervised machine learning algorithm.”
PositionRank: An Unsupervised Approach to Keyphrase Extraction from Scholarly Documents (2017, background): “, 1999) and GenEx (Turney, 2000) are two representative supervised approaches with the most important features being the frequency and the po-”
Automatic Keyphrase Extraction: A Survey of the State of the Art (2014, methodology): “Early supervised approaches to keyphrase extraction recast this task as a binary classiﬁcation problem (Frank et al., 1999; Turney, 1999; Witten et al., 1999; Turney, 2000).”
Graph-Based Keyword Extraction for Single-Document Summarization (2008, methodology): “The supervised learning approach for keywords extraction was ﬁrst suggested in (Turney, 2000), where parametrized heuristic rules were combined with a genetic algorithm into a system - GenEx - that automatically identiﬁed keywords in a document.”