Text Classification for Intelligent

2016

This paper gives a comparison of frequently used classifier models for text classification in the recent years. Text classification is the task of assigning pre-defined categories to free text documents. With the advent of technology we come across a large scale and high dimensional text corpus. It is a challenge to perform text classification on such corpus, in presence of imbalanced class distribution and noisy term features. There are a number of models for performing text classification with various degrees of success. We can simplify the task of text classification by using the following models: Support Vector Machine (SVM), k-nearest neighbor (KNN), Centroid (Rocchio), Naïve Bayes classifiers.

Electronics Letters, 2018

A new algorithm based on learning vector quantisation classifier is presented based on a modified proximity-measure, which enforces a predetermined correct classification level in training while using sliding-mode approach for stable variation in weight updates towards convergence. The proposed algorithm and some well-known counterparts are implemented by using Python libraries and compared in a task of text classification for document categorisation. Results reveal that the new classifier is a successful contender to those algorithms in terms of testing and training performances.

Web Intelligence, 2020

Text classification (a.k.a text categorisation) is an effective and efficient technology for information organisation and management. With the explosion of information resources on the Web and corporate intranets continues to increase, it has being become more and more important and has attracted wide attention from many different research fields. In the literature, many feature selection methods and classification algorithms have been proposed. It also has important applications in the real world. However, the dramatic increase in the availability of massive text data from various sources is creating a number of issues and challenges for text classification such as scalability issues. The purpose of this report is to give an overview of existing text classification technologies for building more reliable text classification applications, to propose a research direction for addressing the challenging problems in text mining.

Book Publisher International (a part of SCIENCEDOMAIN International), 2021

In this paper we present automated text classification in text mining that is gaining greater relevance in various fields every day. Text mining primarily focuses on developing text classification systems able to automatically classify huge volume of documents, comprising of unstructured and semi structured data. The process of retrieval, classification and summarization simplifies extract of information by the user. The finding of the ideal text classifier, feature generator and distinct dominant technique of feature selection leading all other previous research has received attention from researchers of diverse areas as information retrieval, machine learning and the theory of algorithms. To automatically classify and discover patterns from the different types of the documents [1], techniques like Machine Learning, Natural Language Processing (NLP) and Data Mining are applied together. In this paper we review some effective feature selection researches and show the results in a table form.

2013

A new text categorization problem is introduced. As in the classical problem, there is a set of documents and a set of categories. However, in addition to being assigned to a specific category, each document belongs to a certain sequence of documents, referred to as a case. It is assumed that all documents in the same case belong to the same category. An example may be a set of news articles. Their categories may be sport, politics, entertainment, etc. In each category there exist cases, i.e., sequences of documents describing, for example evolution of some events. The problem considered is how to classify a document to a proper category and a proper case within this category. In the paper we formalize the problem and discuss two approaches to its solution.

KSII Transactions on Internet and Information Systems, 2019

The performance of text classification is highly related to the feature selection methods. Usually, two tasks are performed when a feature selection method is applied to construct a feature set; 1) assign score to each feature and 2) select the top-N features. The selection of top-N features in the existing filter-based feature selection methods is biased by their discriminative power and the empirical process which is followed to determine the value of N. In order to improve the text classification performance by presenting a more illustrative feature set, we present an approach via a potent representation learning technique, namely DBN (Deep Belief Network). This algorithm learns via the semantic illustration of documents and uses feature vectors for their formulation. The nodes, iteration, and a number of hidden layers are the main parameters of DBN, which can tune to improve the classifier's performance. The results of experiments indicate the effectiveness of the proposed method to increase the classification performance and aid developers to make effective decisions in certain domains.

This research focuses on Text Classification. Text classification is the task of automatically sorting a set of documents into categories from a predefined set. The domain of this research is the combination of information retrieval (IR) technology, Data mining and machine learning (ML) technology. This research will outline the fundamental traits of the technologies involved. This research uses three text classification algorithms (Naive Bayes, VSM for text classification and the new technique-Use of Stanford Tagger for text classification) to classify documents into different categories, which is trained on two different datasets (20 Newsgroups and New news dataset for five categories).In regards to the above classification strategies, Naïve Bayes is potentially good at serving as a text classification model due to its simplicity.

This paper presents a formal analysis of popular text classification methods, focusing on their loss functions whose minimization is essential to the optimization of those methods, and whose decomposition into the trainingset loss and the model complexity enables cross-method comparisons on a common basis from an optimization point of view. Those methods include Support Vector Machines, Linear Regression, Logistic Regression, Neural Network, Naive Baycs, K-Nearest Neighbor, Rocchio-style and Multi-class Prototype classifiers. Theoretical analysis (including our new derivations) is provided for each method, along with e~-aluation results for all the methods on the Reuters-21578 benchmark corpus. Using linear regression, neural networks and logistic regression methods as examples, we show that properly tuning the balance between the training-set loss and the complexity penalty would have a significant impact to the performance of a classifier. In linear regression, in particular, the tuning of the complexity penalty yielded a result (measured using macro-averaged F1) that outperformed all text categorization methods ever evaluated on that benchmark corpus, including Support Vector Machines.

The printed transformation has seen a gigantic change in the accessibility of online data. Discovering data for pretty much any need has never been more programmed. Content arrangement (otherwise called content classification or point spotting) is the errand of naturally sorting an arrangement of archives into classifications from a predefined set. This assignment has a few applications, including computerized ordering of logical articles, recording licenses into patent indexes, particular spread of data to data purchasers, robotized populace of various leveled inventories of Web assets, spam separating, and recognizable proof of report class. Computerized content characterization is appealing in light of the fact that it liberates associations from the need of physically sorting out report bases, which can be excessively costly, or essentially not plausible since time is running short imperatives of the application or the quantity of records included. The exactness of present day content characterization frameworks equals that of prepared human experts, on account of a blend of information retrieval (IR) innovation and machine learning (ML) innovation. The point of this paper is to highlight the essential calculations that are utilized in content archives grouping, while in the meantime making familiarity with a portion of the fascinating difficulties that stay to be fathomed.

Automatic text classification into predefined categories is an increasingly important task given the vast number of electronic documents available on the Internet and enterprise servers. Successful text classification relies heavily on the vital task of dimensionality reduction, which aims to improve classification accuracy, give greater expression to the classification process, and improve classification computational efficiency. In this paper, two algorithms for feature selection are presented, based on sampling and weighted sampling that build on the C4.5 algorithm. The results demonstrate considerable improvements with regard to classification accuracy - up to 10% - compared to traditional algorithms such as C4.5, Naïve Bayes and Support Vector Machines. The classification process is performed using the Naïve Bayes model in the space of reduced dimensionality. Experiments were carried out using data sets based on the Reuters-21578 collection.