Ana Lorga da Silva

Statistical modelling of the service life prediction of painted surfaces

Service life prediction assumes a primary role as it allows a more rational use of construction elements. This constitutes a useful tool in the definition of preventive maintenance plans providing an increase in performance. The main objective of this research is the development of a statistical methodology for the service life prediction of external painted surfaces. This research is based on field data collected via a survey of the state of deterioration of in-service buildings. The degradation is defined by a number of factors that together contribute to the deterioration of painted surfaces thus ending their service life. In this study a mathematical model was defined using a multiple linear regression analysis and this enables the coatings deterioration over time to be expressed as a function of various degradation factors. 220 painted coatings were inspected in 160 buildings of varying construction types. Analytical tools were devised to monitor the performance of paint coatings on walls and estimate their service life. This study contributes to the automation of the maintenance of painted facades, allowing a more rational management of the maintenance of buildings, converted into economic and performance gains.

Missing Data and Imputation Methods in Partition of Variables

We deal with the effect of missing data under a “Missing at Random Model” on classification of variables with non-hierarchical methods. The partitions are compared by the Rand index.

Missing Data in Hierarchical Classification of Variables — a Simulation Study

Here we develop from a first work the effect of missing data in hierarchical classification of variables according to the following factors: amount of missing data, imputation techniques, similarity coefficient, and aggregation criterion. We have used two methods of imputation, a regression method using an OLS method and an EM algorithm. For the similarity matrices we have used the basic affinity coefficient and the Pearson’s correlation coefficient. As aggregation criteria we apply average linkage, single linkage and complete linkage methods. To compare the structure of the hierarchical classifications the Spearman’s coefficient between the associated ultrametrics has been used. We present here simulation experiments in five multivariate normal cases.

The use of numerical value of adverbs of quantity and frequency in the measurement of behavior patterns: transforming ordinal scales into interval scales

This paper presents a research on rating scales in response to different situations. It aims to improve the significance and accuracy of ordinal scales, transforming them into interval scales. To reach this objective, the presented scales combine quantitative and qualitative perspectives, joining the ease of the Likert scale and the Thurstone’s procedure. In this research, a sample of subjects was asked to indicate the numerical value of adverbs, in reference to a numerical scale. The results were subjected to statistical analysis, to assess their validity. Combining the qualitative dimension with a quantitative evaluation, this procedure can meet the biopsychosocial specificities of subjects, as required by the complexity paradigm. The results of this study seem to be an affirmative response to the questions about validity and reliability, and about the practicality of this procedure.

A lógica complexa da avaliação

This paper presents the theoretical part of a study on the logic which the evaluation process must meet. It explains the concept of general logic and its essential criteria, highlighting the key steps of evaluative reasoning. The authors discuss the problem of the possibility of an objective evaluation, scientifically underpinned. They propose ways to control the factors responsible for some harmful biases to the objectivity of evaluation; define the structural and procedural conditions that affect objectivity; and show that the degree of objectivity depends on the use of an appropriate combination of power and information, the two components of the evaluation process, as a decision making behavior.

The complex logic of evaluation

This paper presents the theoretical part of a study on the logic which the evaluation process must meet. It explains the concept of general logic and its essential criteria, highlighting the key steps of evaluative reasoning. The authors discuss the problem of the possibility of an objective evaluation, scientifically underpinned. They propose ways to control the factors responsible for some harmful biases to the objectivity of evaluation; define the structural and procedural conditions that affect objectivity; and show that the degree of objectivity depends on the use of an appropriate combination of power and information, the two components of the evaluation process, as a decision making behavior.

La lógica compleja de la evaluación

This paper presents the theoretical part of a study on the logic which the evaluation process must meet. It explains the concept of general logic and its essential criteria, highlighting the key steps of evaluative reasoning. The authors discuss the problem of the possibility of an objective evaluation, scientifically underpinned. They propose ways to control the factors responsible for some harmful biases to the objectivity of evaluation; define the structural and procedural conditions that affect objectivity; and show that the degree of objectivity depends on the use of an appropriate combination of power and information, the two components of the evaluation process, as a decision making behavior.