Gilberto Rodrigues Liska

Data classification with binary response through the Boosting algorithm and logistic regression

Abstract The task of classifying is natural to humans, but there are situations in which a person is not best suited to perform this function, which creates the need for automatic methods of classification. Traditional methods, such as logistic regression, are commonly used in this type of situation, but they lack robustness and accuracy. These methods do not not work very well when the data or when there is noise in the data, situations that are common in expert and intelligent systems. Due to the importance and the increasing complexity of problems of this type, there is a need for methods that provide greater accuracy and interpretability of the results. Among these methods, is Boosting, which operates sequentially by applying a classification algorithm to reweighted versions of the training data set. It was recently shown that Boosting may also be viewed as a method for functional estimation. The purpose of the present study was to compare the logistic regressions estimated by the maximum likelihood model (LRMML) and the logistic regression model estimated using the Boosting algorithm, specifically the Binomial Boosting algorithm (LRMBB), and to select the model with the better fit and discrimination capacity in the situation of presence(absence) of a given property (in this case, binary classification). To illustrate this situation, the example used was to classify the presence (absence) of coronary heart disease (CHD) as a function of various biological variables collected from patients. It is shown in the simulations results based on the strength of the indications that the LRMBB model is more appropriate than the LRMML model for the adjustment of data sets with several covariables and noisy data. The following sections report lower values of the information criteria AIC and BIC for the LRMBB model and that the Hosmer–Lemeshow test exhibits no evidence of a bad fit for the LRMBB model. The LRMBB model also presented a higher AUC, sensitivity, specificity and accuracy and lower values of false positives rates and false negatives rates, making it a model with better discrimination power compared to the LRMML model. Based on these results, the logistic model adjusted via the Binomial Boosting algorithm (LRMBB model) is better suited to describe the problem of binary response, because it provides more accurate information regarding the problem considered.

Selecting A Probabilistic Model Applied To The Sensory Analysis Of Specialty Coffees Performed With Consumer

Sensory analysis of cafes assumes that a sensory panel is formed by trained panelists according to recommendations of the American Specialty Coffee Association. However, the choice that determines the preference of a coffee is routinely done through experimentation with consumers, in which largely presents no particular skill in terms of sensory characteristics. Upon this fact, this study aimed to conduct a study considering several probabilistic distributions belonging to the class of generalized extreme value, considering a sensory analysis applied to evaluation of four specialty coffees produced with different processes and at different altitudes in the mountain region of the Mantiqueira state of Minas Gerais. For this analysis, we considered a sensory panel trained to untrained consumers. It was found that the extreme value distribution was the best fit and the final note that the odds of a consumer to submit a maximum score was 9.0 points lower. Therefore, there is evidence to conclude that an efficient identification of specialty coffees produced in this region made by consumers requires more intensive training.

Biomass and Carbon in a Seasonal Semideciduous Forest in Minas Gerais

The objective of the present study was to quantify the biomass and carbon stock in a Seasonal Semideciduous Forest remnant in southern Minas Gerais. Forest inventory data taken between 2010 and 2013 in permanent plots, was used to estimate biomass with an allometric equation. Basic wood density (Db) and carbon content were determined in the laboratory and the carbon stock was obtained by multiplying biomass by carbon content. The species with the lowest and highest Db were Nectandra lanceolata (0.38 g cm-3) and Machaerium villosum (0.77 g cm-3). The species that showed the lowest and highest carbon content values were Casearia decandra (41.85%) and Nectandra oppositifolia (46.57%). The biomass stock for the area was 126.92 ± 0.09 t ha-1, which corresponded to 55.91 ± 0.05 t ha-1 of carbon stock and a periodic annual increment of 3.07 t ha-1 year-1.

Análise Espacial do Processo de Restauração na Região da Usina Hidrelétrica de Camargos, Itutinga-MG

The degradation of natural environments has been a great challenge to be faced when an ecosystem is degraded, damaged, transformed or entirely destroyed as a direct or indirect result of human activities. In this situation, becomes necessary to apply techniques aimed at ecological restoration. To analyze situations like that, ecological indicators are used, among them, the canopy closure, represented by the variable canopy closure index (CCI), it has emerged as a good indicator, as it controls the quantity, quality and the temporal and spatial distribution of light. This study was conducted in the neighborhood of the Camargos Hydropower Plant, the city of Itutinga-MG. Given the above, this study aimed to analyze the CCI variable, in order to make information about their spatial distribution, using resources of geostatistics. The results show the Gaussian model is more appropriate. The estimate for the range was 130.5 meters. It means that until this distance is evidence of spatial dependence statistics on the distribution of CCI. The Ordinary Kriging was used to make the predictions of CCI and the results show the occurrence of several regions with values above 80% of CCI, indicating that the restore process in the region is promising.