Improving Multi-class Text Classification with Naive Bayes

There are numerous text documents available in electronic form. More and more are becoming available every day. Such documents represent a massive amount of information that is easily accessible. Seeking value in this huge collection requires organization; much of the work of organizing documents can be automated through text classification. The accuracy and our understanding of such systems greatly influences their usefulness. In this paper, we seek 1) to advance the understanding of commonly used text classification techniques, and 2) through that understanding, improve the tools that are available for text classification. We begin by clarifying the assumptions made in the derivation of Naive Bayes, noting basic properties and proposing ways for its extension and improvement. Next, we investigate the quality of Naive Bayes parameter estimates and their impact on classification. Our analysis leads to a theorem which gives an explanation for the improvements that can be found in multiclass classification with Naive Bayes using Error-Correcting Output Codes. We use experimental evidence on two commonly-used data sets to exhibit an application of the theorem. Finally, we show fundamental flaws in a commonly-used feature selection algorithm and develop a statistics-based framework for text feature selection. Greater understanding of Naive Bayes and the properties of text allows us to make better use of it in text classification.

A Note on Object Class Representation and Categorical Perception

We present a novel scheme ("Categorical Basis Functions", CBF) for object class representation in the brain and contrast it to the "Chorus of Prototypes" scheme recently proposed by Edelman. The power and flexibility of CBF is demonstrated in two examples. CBF is then applied to investigate the phenomenon of Categorical Perception, in particular the finding by Bulthoff et al. (1998) of categorization of faces by gender without corresponding Categorical Perception. Here, CBF makes predictions that can be tested in a psychophysical experiment. Finally, experiments are suggested to further test CBF.

The Individual is Nothing, the Class Everything: Psychophysics and Modeling of Recognition in Obect Classes

Most psychophysical studies of object recognition have focussed on the recognition and representation of individual objects subjects had previously explicitely been trained on. Correspondingly, modeling studies have often employed a 'grandmother'-type representation where the objects to be recognized were represented by individual units. However, objects in the natural world are commonly members of a class containing a number of visually similar objects, such as faces, for which physiology studies have provided support for a representation based on a sparse population code, which permits generalization from the learned exemplars to novel objects of that class. In this paper, we present results from psychophysical and modeling studies intended to investigate object recognition in natural ('continuous') object classes. In two experiments, subjects were trained to perform subordinate level discrimination in a continuous object class - images of computer-rendered cars - created using a 3D morphing system. By comparing the recognition performance of trained and untrained subjects we could estimate the effects of viewpoint-specific training and infer properties of the object class-specific representation learned as a result of training. We then compared the experimental findings to simulations, building on our recently presented HMAX model of object recognition in cortex, to investigate the computational properties of a population-based object class representation as outlined above. We find experimental evidence, supported by modeling results, that training builds a viewpoint- and class-specific representation that supplements a pre-existing repre-sentation with lower shape discriminability but possibly greater viewpoint invariance.

Visual Recognition and Categorization on the Basis of Similarities to Multiple Class Prototypes

To recognize a previously seen object, the visual system must overcome the variability in the object's appearance caused by factors such as illumination and pose. Developments in computer vision suggest that it may be possible to counter the influence of these factors, by learning to interpolate between stored views of the target object, taken under representative combinations of viewing conditions. Daily life situations, however, typically require categorization, rather than recognition, of objects. Due to the open-ended character both of natural kinds and of artificial categories, categorization cannot rely on interpolation between stored examples. Nonetheless, knowledge of several representative members, or prototypes, of each of the categories of interest can still provide the necessary computational substrate for the categorization of new instances. The resulting representational scheme based on similarities to prototypes appears to be computationally viable, and is readily mapped onto the mechanisms of biological vision revealed by recent psychophysical and physiological studies.

A New Class of Nonionic Photosensitive Surfactants: Some Insights Concerning Conformations

We report on a new class of nonionic, photosensitive surfactants consisting of a polar di(ethylene oxide) head group attached to an alkyl spacer of between two and eight methylene groups, coupled through an ether linkage to an azobenzene moiety. Structural changes associated with the interconversion of the azobenzene group between its cis and trans forms as mediated by the wavelength of an irradiating light source cause changes in the surface tension and self-assembly properties. Differences in saturated surface tensions (surface tension at concentrations above the CMC) were as high as 14.4 mN/m under radiation of different wavelengths. The qualitative behavior of the surfactants changed as the spacer length changed, attributed to the different orientations adopted by the different surfactants depending on their isomerization states, as revealed by neutron reflection studies. The self-assembly of these photosensitive surfactants has been investigated by light scattering, small angle neutron scattering, and cryo-TEM under different illuminations. The significant change in the self-assembly in response to different illumination conditions was attributed to the sign change in Gaussian rigidity, which originated from the azobenzene photoisomerization.

Towards a real-time vision-based navigation system for a small-class UUV

This paper deals with the problem of navigation for an unmanned underwater vehicle (UUV) through image mosaicking. It represents a first step towards a real-time vision-based navigation system for a small-class low-cost UUV. We propose a navigation system composed by: (i) an image mosaicking module which provides velocity estimates; and (ii) an extended Kalman filter based on the hydrodynamic equation of motion, previously identified for this particular UUV. The obtained system is able to estimate the position and velocity of the robot. Moreover, it is able to deal with visual occlusions that usually appear when the sea bottom does not have enough visual features to solve the correspondence problem in a certain area of the trajectory

Class Test Generator

A combination of perl scripts and LaTeX files, this generates multiple multiple choice class tests from a single set of questions. You input a list of questions and answers into a text file. The script then produces any number of class tests that can be used, together with master answer sheets, by scrambling the order of the questions and the answers. Includes a detailed README file, but best just to try it and see.

Collection - class exercise in group presentations with associated materials

slides plus mark scheme and link to group work videos

System Design: UML Class Diagrams

In this session we look at UML Class Diagrams and how they fit into both the family of UML models, and also the software engineering process. We look at some basic features of class diagrams including properties, operations, associations, generalisation, aggregation and composition.

Systems Design: Class modelling using Visual Paradigm

This lab follows the lectures 'System Design: http://www.edshare.soton.ac.uk/6280/ http://www.edshare.soton.ac.uk/9653/ and http://www.edshare.soton.ac.uk/9713/ Students use Visual Paradigm for UML to build Class models through project examples: Aircraft Manufacturing Company, Library, Plant Nursery.

Shared HLA Class II in Six Autoimmune Diseases in Latin America: A Meta-Analysis

The prevalence and genetic susceptibility of autoimmune diseases (ADs) may vary depending on latitudinal gradient and ethnicity. The aims of this study were to identify common human leukocyte antigen (HLA) class II alleles that contribute to susceptibility to six ADs in Latin Americans through a meta-analysis and to review additional clinical, immunological, and genetic characteristics of those ADs sharing HLA alleles. DRB1∗03:01 (OR: 4.04; 95%CI: 1.41–11.53) was found to be a risk factor for systemic lupus erythematosus (SLE), Sjogren’s syndrome (SS), and type 1 diabetes mellitus (T1D). DRB1 ¨ ∗04:05 (OR: 4.64; 95%CI: 2.14–10.05) influences autoimmune hepatitis (AIH), rheumatoid arthritis (RA), and T1D; DRB1∗04:01 (OR: 3.86; 95%CI: 2.32–6.42) is a susceptibility factor for RA and T1D. Opposite associations were found between multiple sclerosis (MS) and T1D. DQB1∗06:02 and DRB1∗15 alleles were risk factors for MS but protective factors for T1D. Likewise, DQB1∗06:03 allele was a risk factor for AIH but a protective one for T1D. Several common autoantibodies and clinical associations as well as additional shared genes have been reported in these ADs, which are reviewed herein. These results indicate that in Latin Americans ADs share major loci and immune characteristics.

Study habits inventory of physical therapy students in a decisions making class

Objective. To present the results of Study Habits Inventory, administered to students of the signature Decision Making IV, of Physiotherapy Program of the Rosario University. Methodology. It was investigated in the second half of 2006, study habits and conditions that influence how students of the subject assumes its university activity. The instrument used was the Inventory of Study Habits - described by Pozar, which let to know how the student engages in his studio, materials, environment and planning in its efforts to exercise their academic role. It was administered to a sample of 46 students enrolled in the 54 subjects at 85% of the total population. Results. 33% of students found that the environmental study is not satisfactory to the learning process, similar to the 41% who said that the planning study is insufficient, the use of materials was evaluated by 61% of students as a positive factor as well as 93% in the category of assimilation of content. This calls to validate the need to confront the meaning and significance for students and professors “content assimilation” as the study shows that it is reverse to academic performance. Discussion It is necessary to implement strategies that strengthen habits in students to achieve academic success and personal training considering the particular conditions of each group of students.

Overcoming Gender Disadvantages. Social Policy Analysis of urban middle-class women in Colombia

The World Bank Report 2012 starts with this statement: “Gender equality matters in itself andit matters for development because, in today’s globalized worlds, countries that use the skillsand talents of their women would have an advantage over those which do not use it.” With theframe that suggest that gender equality matters, this paper describes some policy alternativesoriented to overcome gender disadvantages in the formal labor market incorporation of theurban middle class women in Colombia. On balance, the final recommendation suggest that itis desirable to adopt policy alternatives as Community Centers, which are programs orientedto a social redistribution of the domestic work as a way to encourage women participationin the formal labor market with the social support of the members of their own community.The problem that the social policy needs to address is the segregation of women in the formallabor market in Colombia. Although the evidence shows that the women overcome theeducational gap by showing better performance in education that their male peers, womenare still segregated of the labor market. The persistence of high rates of unemployment on thefemale population, the prevalence of the informal labor market as a women labor market, andthe presence of the payment difference between men and women with similar professionaltrainings are circumstances that sustain the segregation statement. These circumstances areinefficient for the society because an economic analysis shows that the cost of maintain the statuquo is externalized in the social security system that includes health, pension and maternityleave regimens. Therefore, the women segregation involves a market failure.This paper evaluates five policy alternatives each directed to the progress of a different causaldimension of the problem: (i) Quotas in the private market, (ii) Flexible working hours,(iii) replace the maternity leave with a family leave, (iv) Increase the Community Centers forredistributing the care work, and (v) Equal payment enforcement. The first alternative looksto increase women’s participation in the formal labor market. The second, third, and fourthalternatives constitute a package addressed at redistributing care work by reducing women’sresponsibility for reproductive work in the household with the help of husbands and the localgovernment. The fifth alternative intervenes to resolve the equal payment problem.After a four criteria evaluation that measure effectiveness, robustness and improbability inimplementation, efficiency and political acceptability or social opposition, the strongest alternativeis the fostering of Community Centers that promote a redistribution of care work. Thispolicy performs well in the assessment process because it combines gender focus with importantindirect effects: child support and human capabilities. The policy also shows a bottomup implementation process that overcomes the main adoption difficulties in the gender focusprograms and is supported by strong evidence of success in the Colombian context; this evidenceis produced by both transnational actors as a World Bank and also in local accountabilityreporters executed by local institutions like Colombian Institute of Family Welfare (ICBF).

Extra activities for the English class in the first cycle of Primary School. 'Actividades extra para la clase de inglés del primer ciclo de Educación Primaria'.

Resumen basado en el de los autores. Trabajo al que se le concedió una ayuda para la creación de materiales curriculares interactivos en el año 2003

The use of the video camera in the English class. 'La utilización de la cámara de vídeo en la clase de inglés'.

Seminario realizado por cuatro profesoras de distintos centros educativos de Vizcaya para la elaboración de cinco 'projects', cada uno con ejercicios para diferentes niveles, que tienen como objetivo inmediato la práctica de estructuras del lenguaje en un marco comunicativo y como objetivo global la realización de una filmación sobre el tema básico del project. Los temas tratados son: el uso de la cámara de vídeo, las noticias, los anuncios (publicidad), video clips, cuentos y narraciones. Los resultados se valoran como muy positivos debido a la motivación de filmar el producto final.