This paper introduces the idea that if theories of history generate different taxonomies of history they too are incommensurable. I argue this is unavoidable for Kuhn given what he says about incommensurability and 1 ...This paper introduces the idea that if theories of history generate different taxonomies of history they too are incommensurable. I argue this is unavoidable for Kuhn given what he says about incommensurability and 1 investigate the consequences in relation to reflexivity, justification, and paradox for Kuhn＇s account of science. I want to do this on two levels, firstly looking at different possibilities for characterising individual paradigms. I will look at some examples from ancient and early modem astronomy as here it is clearest that paradigms can be characterised in different ways and that this has important consequences. I will argue in particular that Kuhn＇s characterisation of the paradigm for astronomy which emerges from antiquity （geocentrism） is favourable to his general account of the history of science, but that there is a very plausible and extremely damaging alternative. I argue that these differing characterisations generate differing, incommensurable taxonomies of the history of astronomy, with attendant ＂local holism,＂ untranslatability of key terms and issues of theory choice. If so, Kuhn then has problems with generating an adequate decision making protocol for choosing between the two paradigm characterisations. That is problematic in itself, but I also argue this problem is systemic and affects the evidence needed for Kuhn to justify his general account of the history of science. I also want to investigate the implications of differing taxonomies of the history of science at a more abstract level. Kuhn＇s general theory of the history of science generates a taxonomy of the history of science, as do other theories such as those of Popper and of gradualism. If so, the incommensurability involved here, again with attendant ＂local holism,＂ untranslatability of key terms and issues of theory choice, leads to issues of paradox and justification for Kuhn＇s general account of the history of science. With this broader understanding of taxonomic issues, some important Kuhn statements about scientific theories become self-referential, again generating problems of paradox and justification.展开更多
Feature selection is an important approach to dimensionality reduction in the field of text classification. Because of the difficulty in handling the problem that the selected features always contain redundant informa...Feature selection is an important approach to dimensionality reduction in the field of text classification. Because of the difficulty in handling the problem that the selected features always contain redundant information, we propose a new simple feature selection method, which can effectively filter the redundant features. First, to calculate the relationship between two words, the definitions of word frequency based relevance and correlative redundancy are introduced. Furthermore, an optimal feature selection（OFS） method is chosen to obtain a feature subset FS1. Finally, to improve the execution speed, the redundant features in FS1 are filtered by combining a predetermined threshold, and the filtered features are memorized in the linked lists. Experiments are carried out on three datasets（Web KB, 20-Newsgroups, and Reuters-21578） where in support vector machines and na？ve Bayes are used. The results show that the classification accuracy of the proposed method is generally higher than that of typical traditional methods（information gain, improved Gini index, and improved comprehensively measured feature selection） and the OFS methods. Moreover, the proposed method runs faster than typical mutual information-based methods（improved and normalized mutual information-based feature selections, and multilabel feature selection based on maximum dependency and minimum redundancy） while simultaneously ensuring classification accuracy. Statistical results validate the effectiveness of the proposed method in handling redundant information in text classification.展开更多
文摘This paper introduces the idea that if theories of history generate different taxonomies of history they too are incommensurable. I argue this is unavoidable for Kuhn given what he says about incommensurability and 1 investigate the consequences in relation to reflexivity, justification, and paradox for Kuhn＇s account of science. I want to do this on two levels, firstly looking at different possibilities for characterising individual paradigms. I will look at some examples from ancient and early modem astronomy as here it is clearest that paradigms can be characterised in different ways and that this has important consequences. I will argue in particular that Kuhn＇s characterisation of the paradigm for astronomy which emerges from antiquity （geocentrism） is favourable to his general account of the history of science, but that there is a very plausible and extremely damaging alternative. I argue that these differing characterisations generate differing, incommensurable taxonomies of the history of astronomy, with attendant ＂local holism,＂ untranslatability of key terms and issues of theory choice. If so, Kuhn then has problems with generating an adequate decision making protocol for choosing between the two paradigm characterisations. That is problematic in itself, but I also argue this problem is systemic and affects the evidence needed for Kuhn to justify his general account of the history of science. I also want to investigate the implications of differing taxonomies of the history of science at a more abstract level. Kuhn＇s general theory of the history of science generates a taxonomy of the history of science, as do other theories such as those of Popper and of gradualism. If so, the incommensurability involved here, again with attendant ＂local holism,＂ untranslatability of key terms and issues of theory choice, leads to issues of paradox and justification for Kuhn＇s general account of the history of science. With this broader understanding of taxonomic issues, some important Kuhn statements about scientific theories become self-referential, again generating problems of paradox and justification.
基金Project supported by the Joint Funds of the National Natural Science Foundation of China(No.U1509214)the Beijing Natural Science Foundation,China(No.4174105)+1 种基金the Key Projects of National Bureau of Statistics of China(No.2017LZ05)the Discipline Construction Foundation of the Central University of Finance and Economics,China(No.2016XX02)
文摘Feature selection is an important approach to dimensionality reduction in the field of text classification. Because of the difficulty in handling the problem that the selected features always contain redundant information, we propose a new simple feature selection method, which can effectively filter the redundant features. First, to calculate the relationship between two words, the definitions of word frequency based relevance and correlative redundancy are introduced. Furthermore, an optimal feature selection（OFS） method is chosen to obtain a feature subset FS1. Finally, to improve the execution speed, the redundant features in FS1 are filtered by combining a predetermined threshold, and the filtered features are memorized in the linked lists. Experiments are carried out on three datasets（Web KB, 20-Newsgroups, and Reuters-21578） where in support vector machines and na？ve Bayes are used. The results show that the classification accuracy of the proposed method is generally higher than that of typical traditional methods（information gain, improved Gini index, and improved comprehensively measured feature selection） and the OFS methods. Moreover, the proposed method runs faster than typical mutual information-based methods（improved and normalized mutual information-based feature selections, and multilabel feature selection based on maximum dependency and minimum redundancy） while simultaneously ensuring classification accuracy. Statistical results validate the effectiveness of the proposed method in handling redundant information in text classification.
