scholarly journals Extended Pre-Processing Pipeline For Text Classification: On the Role of Meta-Features, Sparsification and Selective Sampling

Author(s):  
Washington Cunha ◽  
Leonardo Rocha ◽  
Marcos A. Gonçalves

Pipelines for Text Classification are sequences of tasks needed to be performed to classify documents. The pre-processing phase of these pipelines involves different ways of manipulating documents for the learning phase. This Master Thesis introduces three new steps into the traditional pre-processing phase: 1) Meta-Features Generation; 2) Sparsification; and 3) Selective Sampling. Our experimental results, based on more than 5.600 measurements, show that our proposal can achieve significant gains in effectiveness when compared to the traditional TF-IDF representation (up to 52%) and word embeddings (up to 46%), at a much lower cost (9.7x faster). Our Master Thesis also includes a thorough and rigorous evaluation of the trade-offs between cost and effectiveness associated with the introduction of these new steps into the pipeline, as well as a comprehensive comparative experimental evaluation of many alternatives. This thesis falls under the topics of (i) Document Management and Classification, (ii) Information Retrieval Models and Techniques, (iii) and Text Database of the SBBD Call for Papers.

2020 ◽  
Vol 57 (4) ◽  
pp. 102263 ◽  
Author(s):  
Washington Cunha ◽  
Sérgio Canuto ◽  
Felipe Viegas ◽  
Thiago Salles ◽  
Christian Gomes ◽  
...  

2017 ◽  
Vol 49 (6) ◽  
pp. 783 ◽  
Author(s):  
Yan WANG ◽  
Zhenchao LIN ◽  
Bowen HOU ◽  
Shijin SUN

1998 ◽  
Vol 37 (6-7) ◽  
pp. 331-336 ◽  
Author(s):  
Stephen Garbaciak ◽  
Philip Spadaro ◽  
Todd Thornburg ◽  
Richard Fox

Sequential risk mitigation approaches the remediation of contaminated sediments in three phases designed to: (1) immediately reduce the ecological and human health risks associated with high levels of contamination, using methods such as the confinement or capping of high-risk materials; (2) reduce the risks associated with moderate levels of pollution to a minimum, on a less urgent schedule and at a lower cost; and (3) address areas of limited contamination through a combination of natural recovery and enhanced natural recovery (to aid or speed those natural processes). Natural recovery, the reduction of contaminant concentrations through natural processes, is based on the practical observation that overall ecosystem recovery appears to be largely a function of time. Sediment decomposition and the mixing of new and old sediments by bottom-dwelling organisms can both contribute to reduced contaminant concentrations. Knowledge of these processes--sediment decomposition, sediment mixing by bottom-dwelling organisms, and chemical residence time is critical in the development of appropriate ecosystem recovery and waste management strategies. Evaluations to support natural recovery predictions are designed to collect and evaluate information necessary to determine whether surface sediment chemical concentrations, with adequate source control, will reach the cleanup standards within a ten-year period.


Author(s):  
Chelsea Barabas

This chapter discusses contemporary debates regarding the use of artificial intelligence as a vehicle for criminal justice reform. It closely examines two general approaches to what has been widely branded as “algorithmic fairness” in criminal law: the development of formal fairness criteria and accuracy measures that illustrate the trade-offs of different algorithmic interventions; and the development of “best practices” and managerialist standards for maintaining a baseline of accuracy, transparency, and validity in these systems. Attempts to render AI-branded tools more accurate by addressing narrow notions of bias miss the deeper methodological and epistemological issues regarding the fairness of these tools. The key question is whether predictive tools reflect and reinforce punitive practices that drive disparate outcomes, and how data regimes interact with the penal ideology to naturalize these practices. The chapter then calls for a radically different understanding of the role and function of the carceral state, as a starting place for re-imagining the role of “AI” as a transformative force in the criminal legal system.


Author(s):  
Maren N. Vitousek ◽  
Laura A. Schoenle

Hormones mediate the expression of life history traits—phenotypic traits that contribute to lifetime fitness (i.e., reproductive timing, growth rate, number and size of offspring). The endocrine system shapes phenotype by organizing tissues during developmental periods and by activating changes in behavior, physiology, and morphology in response to varying physical and social environments. Because hormones can simultaneously regulate many traits (hormonal pleiotropy), they are important mediators of life history trade-offs among growth, reproduction, and survival. This chapter reviews the role of hormones in shaping life histories with an emphasis on developmental plasticity and reversible flexibility in endocrine and life history traits. It also discusses the advantages of studying hormone–behavior interactions from an evolutionary perspective. Recent research in evolutionary endocrinology has provided insight into the heritability of endocrine traits, how selection on hormone systems may influence the evolution of life histories, and the role of hormonal pleiotropy in driving or constraining evolution.


Author(s):  
David Mares

This chapter discusses the role of energy in economic development, the transformation of energy markets, trade in energy resources themselves, and the geopolitical dynamics that result. The transformation of energy markets and their expansion via trade can help or hinder development, depending on the processes behind them and how stakeholders interact. The availability of renewable, climate-friendly sources of energy, domestically and internationally, means that there is no inherent trade-off between economic growth and the use of fossil fuels. The existence of economic, political, social, and geopolitical adjustment costs means that the expansion of international energy markets to incorporate alternatives to oil and coal is a complex balance of environmental trade-offs with no solutions completely free of negative impact risk. An understanding of the supply of and demand for energy must incorporate the institutional context within which they occur, as well as the social and political dynamics of their setting.


2021 ◽  
pp. 108602662199006
Author(s):  
Peter Tashman ◽  
Svetlana Flankova ◽  
Marc van Essen ◽  
Valentina Marano

We meta-analyze research on why firms join voluntary environmental programs (VEPs) to assess the impact of program stringency, or the extent to which they have rigorous, enforceable standards on these decisions. Stringency creates trade-offs for firms by affecting programs’ effectiveness, legitimacy, and adoption costs. Most research considers singular programs and lacks cross program variation needed to analyze program stringency’s impact. Our meta-analysis addresses this by sampling 127 studies and 23 VEPs. We begin by identifying common institutional and resource-based drivers of participation in the literature, and then analyze how program stringency moderates their impacts. Our results suggest that strictly governed VEPs encourage participation among highly visible and profitable firms, and discourage it when informal institutional pressures are higher, and firms have prior experience with other VEPs or quality management standards. We demonstrate that VEP stringency has nuanced effects on firm participation based on the institutional and resource-based factors facing them.


2011 ◽  
Vol 279 (1732) ◽  
pp. 1287-1292 ◽  
Author(s):  
Roi Holzman ◽  
David C. Collar ◽  
Samantha A. Price ◽  
C. Darrin Hulsey ◽  
Robert C. Thomson ◽  
...  

Morphological diversification does not proceed evenly across the organism. Some body parts tend to evolve at higher rates than others, and these rate biases are often attributed to sexual and natural selection or to genetic constraints. We hypothesized that variation in the rates of morphological evolution among body parts could also be related to the performance consequences of the functional systems that make up the body. Specifically, we tested the widely held expectation that the rate of evolution for a trait is negatively correlated with the strength of biomechanical trade-offs to which it is exposed. We quantified the magnitude of trade-offs acting on the morphological components of three feeding-related functional systems in four radiations of teleost fishes. After accounting for differences in the rates of morphological evolution between radiations, we found that traits that contribute more to performance trade-offs tend to evolve more rapidly, contrary to the prediction. While ecological and genetic factors are known to have strong effects on rates of phenotypic evolution, this study highlights the role of the biomechanical architecture of functional systems in biasing the rates and direction of trait evolution.


Sign in / Sign up

Export Citation Format

Share Document