Machine learning in automated text categorization
By Fabrizio Sebastiani. 2001
9,090 citations recorded by the source corpus.
The automated categorization (or classification) of texts into predefined categories has witnessed a booming interest in the last 10 years, due to the increased availability of documents in digital form and the ensuing need to organize them. In the research community the dominant approach to this problem is based on machine learning techniques: a general inductive process automatically builds a classifier by learning, from a set of preclassified documents, the characteristics of the categories. The advantages of this approach over the knowledge engineering approach (consisting in the manual definition of a classifier by domain experts) are a very good effectiveness, considerable savings in terms of expert labor power, and straightforward portability to different domains. This survey discusses the main approaches to text categorization that fall within the machine learning paradigm. We will discuss in detail issues pertaining to three different problems, namely, document representation, classifier construction, and classifier evaluation.
Subfields: Artificial Intelligence, Natural Language Processing
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
- Mohit Bansal
- Percy Liang
- Yoshua Bengio
- Jianfeng Gao
- Graham Neubig
- Luke S. Zettlemoyer
- Quoc V. Le
- Richard Socher
In Natural Language Processing
- Graham Neubig
- Mohit Bansal
- Christopher D. Manning
- Richard Socher
- Percy Liang
- Luke S. Zettlemoyer
- Mirella Lapata
- Yoav Goldberg
Citations in the corpus
- GloVe: Global Vectors for Word Representation (2014, background): “These vectors can be used as features in a variety of applications, such as information retrieval (Manning et al., 2008), document classification (Sebastiani, 2002), question answering (Tellex et al., 2003), named entity recognition (Turian et al., 2010), and parsing (Socher et al., 2013).”
- Do Not Have Enough Data? Deep Learning to the Rescue! (2019, background): “Text classification (Sebastiani 2002), such as classification of emails into spam and not-spam (Shams 2014), is a fundamental research area in machine learning and natu-ral language processing.”
- From Frequency to Meaning: Vector Space Models of Semantics (2010, methodology): “When we classify documents, we are implying that the documents in a class are similar in some way; thus document classification implies some notion of document similarity, and most machine learning approaches to document classification involve a term– document matrix (Sebastiani, 2002).”
- Security Evaluation of Pattern Classifiers under Attack (2014, methodology): “Four feature subsets with size 1,000, 2,000, 10,000 and 20,000 have been selected using the information gain criterion [44].”
- On Security and Sparsity of Linear Classifiers for Adversarial Settings (2016, methodology): “We extract a dictionary of terms (features) from the first 5000 emails (in chronological order) using the same parsing mechanism of SpamAssassin, and then select the 200 most discriminant features according to the information gain criterion [23].”
- Adversarial Feature Selection Against Evasion Attacks (2016, methodology): “reduced the feature set from more than 20,000 to 500 features, without significant loss in classification accuracy, using a supervised feature selection approach based on the information gain criterion [68].3 The linear SVM was considered as the classification algorithm. As for the performance measure ”
- Text Categorization as a Graph Classification Problem (2015, background): “We refer to (Sebastiani, 2002) for an in-depth review of the earliest works in the field and (Aggarwal and Zhai, 2012) for a survey of the more recent works that capitalize on additional metainformation.”
- Authorship verification as a one-class classification problem (2004, background): “If it can be assumed for each test document that one of the specified authors is indeed the actual author, the problem fits the standard paradigm of a text categorization problem (Sebastiani 2002).”
- Machine Translation Detection from Monolingual Web-Text (2013, methodology): “Although there are several approaches for feature selection, e.g ., (Sebastiani, 2002), we use a method that is suitable for handling a large number of feature candidates.”
- Support Vector Machine Active Learning with Applications to Text Classification (2000, methodology): “Th ere have been many approaches to achieve this goal (Rocchio, 1971; Dumais et al ., 1998; Sebastiani, 2001).”
- Terms of a Feather: Content-Based News Recommendation and Discovery Using Twitter (2011, methodology): “Each term, qi, is used as a query against the RSS index to retrieve the set of articles A that contain qi along with their associated TF-IDF (term frequency inverse document frequency) score [16].”
- Enhancing the LexVec Distributed Word Representation Model Using Positional Contexts and External Memory (2016, background): “Distributed word representations have become a mainstay innatural language processing , enjoying a slew of applications (Sebastiani, 2002; Turian et al., 2010; Socher et al., 2013).”
- Supervised term weighting for automated text categorization (2003, background): “Text categorization (TC) is the activity of automatically building, by means of machine learning (ML) techniques, automatic text classifiers , i.e. programs capable of labelling natural language texts from a domain D with thematic categories from a predefined set C = { c 1 , . . . , c |C| } [10].”
- Learning to classify short and sparse text & web with hidden topics from large-scale data collections (2008, positive): “) and achieved satisfactory results [3, 32].”
- Social Media Analytics (2014, background): “Recently a number of novel approaches made their way into the social sciences, most notably text classification based on unsupervised and supervised learning (Sebastiani 2002; Liu 2011).”
- A systematic analysis of performance measures for classification tasks (2009, background): “It is dedicated to finding texts according to a given criteria (Sebastiani, 2002) and it includes the classification of documents (research papers, technical reports, magazine articles, etc.).”
- TEXT CLASSIFICATION AND CLASSIFIERS: A SURVEY (2012, methodology): “Although for reasons of space we will not discuss them in detail, we at least want to mention the existence of WORD, Sleeping expert, CONSTRUE [54], genetic, Online Classifier [4] Fuzzy correlation and some Hybrid technique are given in [3].”