Evolutionary model trees for handling continuous classes in machine learning

Model trees are a particular case of decision trees employed to solve regression problems. They have the advantage of presenting an interpretable output, helping the end-user to get more confidence in the prediction and providing the basis for the end-user to have new insight about the data, confirming or rejecting hypotheses previously formed. Moreover, model trees present an acceptable level of predictive performance in comparison to most techniques used for solving regression problems. Since generating the optimal model tree is an NP-Complete problem, traditional model tree induction algorithms make use of a greedy top-down divide-and-conquer strategy, which may not converge to the global optimal solution. In this paper, we propose a novel algorithm based on the use of the evolutionary algorithms paradigm as an alternate heuristic to generate model trees in order to improve the convergence to globally near-optimal solutions. We call our new approach evolutionary model tree induction (E-Motion). We test its predictive performance using public UCI data sets, and we compare the results to traditional greedy regression/model trees induction algorithms, as well as to other evolutionary approaches. Results show that our method presents a good trade-off between predictive performance and model comprehensibility, which may be crucial in many machine learning applications. (C) 2010 Elsevier Inc. All rights reserved.

Fragment-guided approach to incorporating structural information into a CoMFA study: BACE-1 as an example

Alzheimer`s disease is an ultimately fatal neurodegenerative disease, and BACE-1 has become an attractive validated target for its therapy, with more than a hundred crystal structures deposited in the PDB. In the present study, we present a new methodology that integrates ligand-based methods with structural information derived from the receptor. 128 BACE-1 inhibitors recently disclosed by GlaxoSmithKline R&D were selected specifically because the crystal structures of 9 of these compounds complexed to BACE-1, as well as five closely related analogs, have been made available. A new fragment-guided approach was designed to incorporate this wealth of structural information into a CoMFA study, and the methodology was systematically compared to other popular approaches, such as docking, for generating a molecular alignment. The influence of the partial charges calculation method was also analyzed. Several consistent and predictive models are reported, including one with r (2) = 0.88, q (2) = 0.69 and r (pred) (2) = 0.72. The models obtained with the new methodology performed consistently better than those obtained by other methodologies, particularly in terms of external predictive power. The visual analyses of the contour maps in the context of the enzyme drew attention to a number of possible opportunities for the development of analogs with improved potency. These results suggest that 3D-QSAR studies may benefit from the additional structural information added by the presented methodology.

University Scheduling using Genetic Algorithm

The automated timetabling and scheduling is one of the hardest problem areas. This isbecause of constraints and satisfying those constraints to get the feasible and optimizedschedule, and it is already proved as an NP Complete (1) [1]. The basic idea behind this studyis to investigate the performance of Genetic Algorithm on general scheduling problem underpredefined constraints and check the validity of results, and then having comparative analysiswith other available approaches like Tabu search, simulated annealing, direct and indirectheuristics [2] and expert system. It is observed that Genetic Algorithm is good solutiontechnique for solving such problems and later analysis will prove this argument. The programis written in C++ and analysis is done by using variation in various parameters.

Finding rule groups to classify high dimensional gene expression datasets

Microarray data provides quantitative information about the transcription profile of cells. To analyze microarray datasets, methodology of machine learning has increasingly attracted bioinformatics researchers. Some approaches of machine learning are widely used to classify and mine biological datasets. However, many gene expression datasets are extremely high dimensionality, traditional machine learning methods can not be applied effectively and efficiently. This paper proposes a robust algorithm to find out rule groups to classify gene expression datasets. Unlike the most classification algorithms, which select dimensions (genes) heuristically to form rules groups to identify classes such as cancerous and normal tissues, our algorithm guarantees finding out best-k dimensions (genes), which are most discriminative to classify samples in different classes, to form rule groups for the classification of expression datasets. Our experiments show that the rule groups obtained by our algorithm have higher accuracy than that of other classification approaches

Touch it, feel it and experience it: developing professional IS skills using interview-style experiential simulations

The IS education field has made increasing use of computerised experiential simulations, but few attempts have been made to create an authentic learning environment that combines and balances elements of video-based computer simulation with real-life learning activities. This paper explores the design principles used to develop a CD-ROM simulation where learners use interviewing skills to elicit system requirements from simulated employees in an authentic context. The employees are videoed actors who converse with each other and with learners within a dynamic interaction model. The paper also describes how we combined this simulation with other teaching approaches such as in-class discussions, student team work, formal presentations, etc.

Linking attitudes and demographics in a tourist segmentation model : a two-stage approach

Segmentation has been widely studied in tourism research e.g. Dolnicar (2004). Dawley (2006) points that commonly used segmentation variables such as demographics lead to identifiable segments which are not actionable while other useful approaches e.g. psychographics, are actionable but not identifiable. The objective of this paper is to develop a two-stage linkage approach to segmentation whereby cluster analysis using psychographic variables is conducted within demographic group. Demographic groups are selected based on propensity to travel. This research utilizes data generated from a cross-sectional self-completed survey of 49,105 Australian respondents on travel and tourism. The managerial usefulness of this segmentation is assessed. Clearly segments can be directly linked both demographically and psychographically.

Finding rule groups to classify high dimensional gene expression datasets

Microarray data provides quantitative information about the transcription profile of cells. To analyse microarray datasets, methodology of machine learning has increasingly attracted bioinformatics researchers. Some approaches of machine learning are widely used to classify and mine biological datasets. However, many gene expression datasets are extremely high dimensionality, traditional machine learning methods cannot be applied effectively and efficiently. This paper proposes a robust algorithm to find out rule groups to classify gene expression datasets. Unlike the most classification algorithms, which select dimensions (genes) heuristically to form rules groups to identify classes such as cancerous and normal tissues, our algorithm guarantees finding out best-k dimensions (genes) to form rule groups for the classification of expression datasets. Our experiments show that the rule groups obtained by our algorithm have higher accuracy than that of other classification approaches.

Semantic and layered protein function prediction from PPI networks

Background The past few years have seen a rapid development in novel high-throughput technologies that have created large-scale data on protein-protein interactions (PPI) across human and most model species. This data is commonly represented as networks, with nodes representing proteins and edges representing the PPIs. A fundamental challenge to bioinformatics is how to interpret this wealth of data to elucidate the interaction of patterns and the biological characteristics of the proteins. One significant purpose of this interpretation is to predict unknown protein functions. Although many approaches have been proposed in recent years, the challenge still remains how to reasonably and precisely measure the functional similarities between proteins to improve the prediction effectiveness.

Results We used a Semantic and Layered Protein Function Prediction (SLPFP) framework to more effectively predict unknown protein functions at different functional levels. The framework relies on a new protein similarity measurement and a clustering-based protein function prediction algorithm. The new protein similarity measurement incorporates the topological structure of the PPI network, as well as the protein's semantic information in terms of known protein functions at different functional layers. Experiments on real PPI datasets were conducted to evaluate the effectiveness of the proposed framework in predicting unknown protein functions.

Conclusion The proposed framework has a higher prediction accuracy compared with other similar approaches. The prediction results are stable even for a large number of proteins. Furthermore, the framework is able to predict unknown functions at different functional layers within the Munich Information Center for Protein Sequence (MIPS) hierarchical functional scheme. The experimental results demonstrated that the new protein similarity measurement reflects more reasonably and precisely relationships between proteins.

Reducing unhealthy weight gain in Fijian adolescents : results of the Healthy Youth Healthy Communities study

Obesity is a significant problem among adolescents in Pacific populations. This paper reports on the outcomes of a 3-year obesity prevention study, Healthy Youth Healthy Communities, which was part of the Pacific Obesity Prevention in Communities project, undertaken with Fijian adolescents. The intervention was developed with schools and comprised social marketing, nutrition and physical activity initiatives and capacity building designed to reduce unhealthy weight, and the individual exposure period was just over 2-year duration. The evaluation incorporated a quasi-experimental, longitudinal design in seven intervention secondary schools near Suva (n = 874) and a matched sample of 11 comparison secondary schools from western Viti Levu (n = 2,062). There were significant differences between groups at baseline; the intervention group was shorter, weighed less, had a higher proportion of underweight and lower proportion of overweight, and better quality of life (Pediatric Quality of Life Inventory only). At follow-up, the intervention group had lower percentage body fat (-1.17) but also a lower increase in quality of life (Assessment of Quality of Life instrument: -0.02; Pediatric Quality of Life Inventory: -1.94) than the comparison group. There were no other differences in anthropometry, and behaviours’ changes showed a mixed pattern. In conclusion, this school-based health promotion programme lowered percentage body fat but did not reduce unhealthy weight gain or influence most obesity-promoting behaviours among Fijian adolescents. Despite growing evidence supporting the efficacy of community-based approaches to reduce obesity among children of European descent, findings from this study failed to demonstrate the efficacy of a community capacity-building approach among an adolescent sample drawn from a different sociocultural, economic and geographical context. Additional ‘top–down’ or other innovative approaches may be needed to reduce adolescent obesity in the Pacific.

Bias and precision of measures of survival gain from right-censored data

In cost-effectiveness analyses of drugs or health technologies, estimates of life years saved or quality-adjusted life years saved are required. Randomised controlled trials can provide an estimate of the average treatment effect; for survival data, the treatment effect is the difference in mean survival. However, typically not all patients will have reached the endpoint of interest at the close-out of a trial, making it difficult to estimate the difference in mean survival. In this situation, it is common to report the more readily estimable difference in median survival. Alternative approaches to estimating the mean have also been proposed. We conducted a simulation study to investigate the bias and precision of the three most commonly used sample measures of absolute survival gain - difference in median, restricted mean and extended mean survival - when used as estimates of the true mean difference, under different censoring proportions, while assuming a range of survival patterns, represented by Weibull survival distributions with constant, increasing and decreasing hazards. Our study showed that the three commonly used methods tended to underestimate the true treatment effect; consequently, the incremental cost-effectiveness ratio (ICER) would be overestimated. Of the three methods, the least biased is the extended mean survival, which perhaps should be used as the point estimate of the treatment effect to be inputted into the ICER, while the other two approaches could be used in sensitivity analyses. More work on the trade-offs between simple extrapolation using the exponential distribution and more complicated extrapolation using other methods would be valuable.

Retrospective analysis of cancer survival across South-Western Victoria in Australia

OBJECTIVE: This paper aims to describe cancer survival and examine association between survival and socio-demographic characteristics across Barwon South-Western region (BSWR) in Victoria, Australia. DESIGN: This study is based on the retrospective cohort database of patients accessing oncology services across BSWR. SETTING: Six rural and three urban hospital settings across the BSWR. PARTICIPANTS: The participants were patients who were diagnosed with cancer in 2009. MAIN OUTCOME MEASURES: Overall survival (OS) of participants was the main outcome measure. RESULTS: Total of 1778 eligible patients had four-year OS for all cancers combined of 59.7% (95% CI, 57.4-62.0). Improved OS was observed for patients in the upper socio-economic tertile (64.2%; 95% CI, 60.9-67.5) compared to the middle (59.3%; 95% CI, 55.5-63.1) and lowest tertiles (49.6%; 95% CI, 44.2-54.9) (P < 0.01). On multivariate analyses, higher socio-economic status remained a significant predictor of OS adjusting for gender, remoteness and age (HR [hazard ratio] 0.81; 95% CI 0.74-0.89; P < 0.01). Remoteness was significantly associated with improved OS after adjusting for age, gender and socio-economic status (HR 0.86; 95% CI, 0.77-0.97; P = 0.01). Older age ≥70 years compared to <70 years conferred inferior OS (HR 3.08; 95% CI, 2.64-3.59; P < 0.01). CONCLUSIONS: Our study confirmed improved survival outcomes for patients of higher socio-economic status and younger age. Future research to explain the unexpected survival benefit in patients who lived in more remote areas should examine factors including the correlation between geographical residence and eventual treatment facility as well as compare the BSWR care model to other regions' approaches.

Testing the focal species approach to making conservation decisions for species persistence

Aim: Most risk assessments and decisions in conservation are based on surrogate approaches, where a group of species or environmental indicators are selected as proxies for other aspects of biodiversity. In the focal species approach, a suite of species is selected based on life history characteristics, such as dispersal limitation and area requirements. Testing the validity of the focal species concept has proved difficult, due to a lack of theory justifying the underlying framework, explicit objectives and measures of success. We sought to understand the conditions under which the focal species concept has merit for conservation decisions. Location: Our model system comprised 10 vertebrate species in 39 patches of native forest embedded in pine plantation in New South Wales, Australia. Methods: We selected three focal species based on ecological traits. We used a multiple-species reserve selection method that minimizes the expected loss of species, by estimating the risk of extinction with a metapopulation model. We found optimal reserve solutions for multiple species, including all 10 species, the three focal species, for all possible combinations of three species, and for each species individually. Results: Our case study suggests that the focal species approach can work: the reserve system that minimized the expected loss of the focal species also minimized the expected species loss in the larger set of 10 species. How well the solution would perform for other species and given landscape dynamics remains unknown. Main conclusions: The focal species approach may have merit as a conservation short cut if placed within a quantitative decision-making framework, where the aspects of biodiversity for which the focal species act as proxies are explicitly defined, and success is determined by whether the use of the proxy results in the same decision. Our methods provide a framework for testing other surrogate approaches used in conservation decision-making and risk assessment. © 2013 John Wiley & Sons Ltd.

Anisotropy for metals in non-associated flow plasticity

 A constitutive model based on Non-Associated Flow rule is implemented numerically and is shown to be capable of accurate predictions of anisotropy driven phenomena, observed during the forming processes of thin sheet metals, in a more efficient manner than other traditional approaches based on Associated Flow Rule.

Projeto de reestruturação organizacional de uma empresa de construção civil : um projeto de consultoria

The managerial diagnostic carried out at Construtora e Incorporadora A. Bueno Ltda. ¿ CIAB, points to an administrative structure within the concepts of the Classic Theory of Administration. We believe that, to the solution of its operations, it is necessary to turn to the main principle of the Theory of Administration (preview, organize, command, coordinate and control) as a way of finding, according to Taylor and Fayol studies, one possible solution to this case. Nevertheless, the companies adopt other administrative approaches on its operational process although, we understand that nothing has changed, they are still dependent and/or are connected to the principles of these classic authors. So, to the solution of the internal disorganized problems, it is priority the introduction of administrative processes and also the restructuring of the operational processes such as: the study of the time and movement, the rationalization of the physical and material efforts, the supplement of objectives and profitability. Then, according to Fayol the administration is not an exclusive privilege and a personal responsibility of the boss or the leaders of the company; it is a function that shares, like other essential functions, among the head and the members of the company. Consequently, to Fayol, the administration is the search of the maxim prosperity to all involved. So, this study has the aim to elaborate a consultant project to the organizational structurate of the CIAB company, according to the Theory of the Administration.

A formação dos limites no desenvolvimento do indivíduo

Este trabalho foi realizado com base, principalmente, nas contribuições de autores psicanalíticos que, na abordagem do processo de formação da personalidade, enfatizam as primeiras relações que o indivíduo estabelece no seu ambiente imediato, especialmente, com a 'pessoa maternal'. Por outro lado, enfoques que podem ser considerados complementares, são também utilizados, na medida em que auxiliam a compreensão dos complexos fatores envolvidos na formação das fronteiras individuais. São, desse modo, focalizados, em duas etapas fundamentais (e principais), os processos que, no desenvolvimento normal, levam, a partir de um estado geral de indiferenciação, à distinção entre 'EU' e o 'OUTRO' e a um resultante sentimento de identidade pessoal. São, também, abordados os desenvolvimentos não satisfatórios e suas prováveis implicações nos distúrbios psicopatológicos posteriores. A importância da 'pessoa maternal' é destacada por sua ativa participação no progresso da criança 'rumo à independência'. Além de suprir 'suficientemente bem' as suas necessidades, ela deve, amorosamente, permitir à criança vivenciar a si mesma como um ser 'real', para que ela possa alcançar o sentimento do 'EU'. Os limites do indivíduo podem ser, finalmente, visualizados, não apenas em termos de uma dimensão espacial (limites físicos) e de uma dimensão temporal (continuidade de ser), mas, sobretudo, em termos de 'uma dimensão relacional. É neste campo que o indivíduo pode realizar urna diferenciação genuína, ou tornar-se um reflexo das diferenciações de outros.