Reducing performance bias for unbalanced text mining

In text categorization applications, class imbalance, which refers to an uneven data distribution where one class is represented by far more less instances than the others, is a commonly encountered problem. In such a situation, conventional classifiers tend to have a strong performance bias, which results in high accuracy rate on the majority class but very low rate on the minorities. An extreme strategy for unbalanced, learning is to discard the majority instances and apply one-class classification to the minority class. However, this could easily cause another type of bias, which increases the accuracy rate on minorities by sacrificing the majorities. This paper aims to investigate approaches that reduce these two types of performance bias and improve the reliability of discovered classification rules. Experimental results show that the inexact field learning method and parameter optimized one-class classifiers achieve more balanced performance than the standard approaches.

Boys and girls talking about what matters: student voice as text in the English classroom

For some years now we have been talking with young people across Australia. They have shared their experiences with us about school, family, their friends, relationships and just life in general (see Pallotta-Chiarolli 1998, Martino & Pallotta-Chiarolli 200la). Our major aim in this work has been to give young people the opportunity to 'speak their hearts and minds', to collaborate with us in the structuring and stylisation of a text 'by them and for them', and to enable their voices to be heard in the broader society, beyond the exclusive space of the academic journal (see Le Compte 1993). This is established praxis in feminist and postcolonial research that challenges the detached and hierarchical relations between researcher and researched in traditional Western masculinist research.

From site to text : Australian Aborigines and the origin of the family

Missions were not simply sites of modernity, they were also the source of key data for the modernist theories of human progress. The idea that so called “primitive peoples” provided a window to the origins of human institutions seemed axiomatic to nineteenth-century theorists of human society who sought evidence for these ideas from settlers, administrators and particularly missionaries. The 1870s and 1880s were the high point of missionary engagement with study-bound anthropologists, as questionnaires and letters were sent from the centres to the edges of empires. Missionary responses, augmented with settler and explorer observations, became the footnotes in early anthropological texts on “primitive” societies. These analyses were then mined for the foundation texts of the other social sciences in the late nineteenth century. Along with many other scholars, Karl Marx and Friedrich Engels read the anthropology of the period and slotted the findings into their analyses of human society.

Logic and The Open Society : revising the place of Tarski’s theory of truth within Popper’s political philosophy

This chapter retraces the way in which the Austrian philosopher Sir Karl Popper came to accept a Correspondence Theory of Truth from the work of the Polish logician and mathematician Alfred Tarski. It is argued that Popper’s use of Tarski’s semantic theory of truth reveals crucial insights into the fundamental characteristics of Popper’s social philosophy.  Quite deceptively, arguments based upon Tarski’s theory of truth appear implicitly throughout the text of The Open Society and Its Enemies (1945). It is then demonstrated how Popper integrated a correspondence theory of truth into a theory of the functions of communicative language that he received from Karl Bühler.