Estudo Comparativo de Métricas de Pontuação para Aprendizagem Estrutural de Redes Bayesianas

Bayesian networks are powerful tools as they represent probability distributions as graphs. They work with uncertainties of real systems. Since last decade there is a special interest in learning network structures from data. However learning the best network structure is a NP-Hard problem, so many heuristics algorithms to generate network structures from data were created. Many of these algorithms use score metrics to generate the network model. This thesis compare three of most used score metrics. The K-2 algorithm and two pattern benchmarks, ASIA and ALARM, were used to carry out the comparison. Results show that score metrics with hyperparameters that strength the tendency to select simpler network structures are better than score metrics with weaker tendency to select simpler network structures for both metrics (Heckerman-Geiger and modified MDL). Heckerman-Geiger Bayesian score metric works better than MDL with large datasets and MDL works better than Heckerman-Geiger with small datasets. The modified MDL gives similar results to Heckerman-Geiger for large datasets and close results to MDL for small datasets with stronger tendency to select simpler network structures

Versão breve da social rhythm metric: um instrumento para avaliar ritmo social de pacientes com acidente vascular cerebral

Stroke is a neurological disorder caused by restriction of blood flow to the brain, which generates directly a deficit of functionality that affects the quality of life of patients. The aim of this study was to establish a short version of the Social Rhythm Scale (SRM), to assess the social rhythm of stroke patients. The sample consisted of 84 patients, of both sexes, with injury time exceeding 6 months. For seven days, patients recorded the time held 17 activities of SRM. Data analysis was performed using a principal components factor analysis with varimax rotation of the full version of SRM in order to determine which activities could compose brief versions of SRM. We then carried out a comparison of hits, the ALI (Level Activity Index) and SRM, between versions, by Kruskal-Walls and the Mann-Whitney test. The Spearman correlation test was used to evaluate the correlation between the score of the full version of SRM with short versions. It was found that the activities of SRM were distributed in three versions: the first and second with 6 activities and third with 3 activities. Regarding hits, it was found that they ranged from 4.9 to 5.8 on the first version; 2.3 to 3.8 in version 2 and 2.8 to 6.2 in version 3, the first the only version that did not show low values. The analysis of ALI, in version 1, the median was 29, in version 2 was 14 and in version 3 was 18. Significant difference in the values of ALI between versions 1 and 2, between 2 and 3 and between versions 1 and 3. The highest median was found in the first version, formed by activities: out of bed, first contact, drink coffee, watch TV in the evening and go to bed. The lowest median was observed in the second version and this was not what had fewer activities, but which had social activities. The medians of the SRM version 1 was 6, version 2 was 4 and version 3 was 6. Significant difference in the values of SRM between versions 1 and 2 and between 2 and 3, but no significant difference between versions 1 and 3. Through analysis, we found a significant correlation only between the full version and the version 1 (R2 = 0.61) (p <0.05), no correlation was found with version 2 (R2 = 0.007) nor with version 3 (R2 = 0.002), this was finally a factor to consider version 1 as the short brazilian version of the Social Rhythm Metric for stroke patients

Nursing Activities Score: avaliação da carga de trabalho de enfermagem em unidade de terapia intensiva adulto

The sizing of nursing human resources is an essential management tool to meet the needs of the patients and the institution. Regarding to the Intensive Care Unit, where the most critical patients are treated and the most advanced life-support equipments are used, requiring a high number of skilled workers, the use of specific indicators to measure the workload of the team becomes necessary. The Nursing Activities Score is a validated instrument for measuring nursing workload in the Intensive Care Unit that has demonstrated effectiveness. It is a cross-sectional study with the primary objective of assessing the workload of nursing staff in an adult Intensive Care Unit through the application of the Nursing Activities Score. The study was conducted in a private hospital specialized in the treatment of patients with cancer, which is located in the city of Natal (Rio Grande do Norte – Brazil). The study was approved by the Research Ethics Committee of the hospital (Protocol number 558.799; CAAE 24966013.7.0000.5293). For data collection, a form of sociodemographic characteristics of the patients was used; the Nursing Activities Score was used to identify the workload of nursing staff; and the instrument of Perroca, which classifies patients and provides data related to the their need for nursing care, was also used. The collected data were analyzed using a statistical package. The categorical variables were described by absolute and relative frequency, while the number by median and interquartile range. Considering the inferential approach, the Spearman test, the Wald chi-square, Kruskal Wallis and Mann-Whitney test were used. The statistically significant variables were those with p values <0.05. The evaluation of the overall averages of NAS, considering the first 15 days of hospitalization, was performed by the analysis of Generalized Estimating Equations (GEE), with adjust for the variable length of hospitalization. The sample consisted of 40 patients, in the period of June to August 2014. The results showed a mean age of 62,1 years (±23,4) with a female predominance (57,5%). The most frequent type of treatment was clinical (60,0%), observing an average stay of 6,9 days (±6,5). Considering the origin, most patients (35%) came from the Surgical Center. There was a mortality rate of 27,5%. 277 measures of NAS score and Perroca were performed, and the averages of 69,8% (±24,1) and 22,7% (±4.2) were obtained, respectively. There was an association between clinical outcome and value of the Nursing Activities Score in 24 hours (p <0.001), and between the degree of dependency of patients and nursing workload (rp 0,653, p<0,001). The achieved workload of the nursing staff, in the analyzed period, was presented high, showing that hospitalized patients required a high demand for care. These findings create subsidies for sizing of staff and allocation of human resources in the sector, in order to achieve greater safety and patient satisfaction as a result of intensive care, as well as an environment conducive to quality of life for the professionals