Exploiting temporal contexts in text classification

Abstract

Due to the increasing amount of information being stored and accessible through the Web, Automatic Document Classification (ADC) has become an important research topic. ADC usually employs a supervised learning strategy, where we first build a classification model using pre-classified documents and then use it to classify unseen documents. One major challenge in building classifiers is dealing with the temporal evolution of the characteristics of the documents and the classes to which they belong. However, most of the current techniques for ADC do not consider this evolution while building and using the models. Previous results show that the performance of classifiers may be affected by three different temporal effects (class distribution, term distribution and class similarity). Further, it is shown that using just portions of the pre-classified documents, which we call contexts, for building the classifiers, result in better performance, as a consequence of the minimization of the aforementioned effects.In this paper we define the concept of temporal contexts as being the portions of documents that minimize those effects. We then propose a general algorithm for determining such contexts, discuss its implementation-related issues, and propose a heuristic that is able to determine temporal contexts efficiently. In order to demonstrate the effectiveness of our strategy, we evaluated it using two distinct collections: ACM-DL and MedLine. We initially evaluated the reduction in terms of both the effort to build a classifier and the entropy associated with each context. Further, we evaluated whether these observed reductions translate into better classification performance by employing a very simple classifier, majority voting. The results show that we achieved precision gains of up to 30% compared to a version that is not temporally contextualized, and the same accuracy of a state-of-the-art classifier (SVM), while presenting an execution time up to hundreds of times faster.

A Technique for Improving the Performance of Naive Bayes Text Classification

“…Our proposed method transforms to estimate the probability of a given feature belonging to a category, which can be calculated both in a global view and a local view, we will explain it later. In (6), P (c j |w k ) reflects the relationship between features and categories , that is how likely a particular feature given belongs to a category, it is somewhat similar to the concept of dominance which is used in a temporal weighting function mentioned by L. Rocha et al [10]. The normal way to calculate P (c j |w k ) is:…”

On Efficient Meta-Level Features for Effective Text Classification

“…This dataset was obtained from [19]. In that work the authors considered the first level of the taxonomy so that each document article is classified under only one category, avoiding dealing with multilabel cases.…”

Improving tweet stream classification by detecting changes in word probability

“…. ., a classifier is given a new training document set, Temporally-aware classification methods for batch learning have also been proposed [21,5,28,29]: a classifier is given D * t and Dt for all t = 1, 2, . .…”

“…Instance weighting methods assume that newer instances are more essential for classifying the current instance [15,18,5,11,28,29]. Lebanon et al proposed a naive Bayes classifier with a weighting scheme for handling temporal document streams [18].…”