Learning in imbalanced relational data

Traditional learning techniques learn from flat data files with the assumption that each class has a similar number of examples. However, the majority of real-world data are stored as relational systems with imbalanced data distribution, where one class of data is over-represented as compared with other classes. We propose to extend a relational learning technique called Probabilistic Relational Models (PRMs) to deal with the imbalanced class problem. We address learning from imbalanced relational data using an ensemble of PRMs and propose a new model: the PRMs-IM. We show the performance of PRMs-IM on a real university relational database to identify students at risk.

Multi-class pattern classification in imbalanced data

The majority of multi-class pattern classification techniques are proposed for learning from balanced datasets. However, in several real-world domains, the datasets have imbalanced data distribution, where some classes of data may have few training examples compared for other classes. In this paper we present our research in learning from imbalanced multi-class data and propose a new approach, named Multi-IM, to deal with this problem. Multi-IM derives its fundamentals from the probabilistic relational technique (PRMs-IM), designed for learning from imbalanced relational data for the two-class problem. Multi-IM extends PRMs-IM to a generalized framework for multi-class imbalanced learning for both relational and non-relational domains.

An empirical investigation of high relational orientation sport club members

Relationships between businesses, businesses and end customers, as well as between customers are an important area of practical and scientific interest. In the present era, largely due to digital technologies such as the database, public and private networks, and data collection and information distribution via TCP/IP (Transmission Control Protocol/Internet Protocol) interface tools such as the World Wide Web (Web), the interest in relationships and related aspects such as trust as it relates to Web interactivity continues. An important antecedent empirical study established that arts patrons (customers) of a New York theatre company could be segmented according to their relational orientation, and that this orientation mediated between component attitudes and future purchase intentions. The study reported in this paper employs Web-based data collection and postal data collection methods in an investigation of the mediation effects of these data collection methods used with the same population of a premier football club in Australia. While a future aim is to more closely compare the outcomes established in the arts study with those from a similarly constructed study in the sporting arena, the focus of this initial paper is the differences in response exhibited by online respondents relative to postal survey respondents. The paper reports findings which do not support those of the antecedent arts and entertainment study concerning the weakness of overall satisfaction on the purchase intentions of high relational orientation customers. The paper also reports findings which give confidence to users of online surveys that despite differences in demographic profiles of these respondents and postal survey respondents, there is a degree of similarity in the responses of the two groups on the measures used in this study. The paper also suggests the need for further research into these data collection effects as they relate to relationship marketing.

Pubertal stage and the prevalence of violence and social/relational aggression

OBJECTIVE We examined associations between pubertal stage and violent adolescent behavior and social/relational aggression.

METHODS The International Youth Development Study comprises statewide representative student samples in grades 5, 7, and 9 (N = 5769) in Washington State and Victoria, Australia, drawn as a 2-stage cluster sample in each state. We used a school-administered, self-report student survey to measure previous-year violent behavior (ie, attacking or beating up another person) and social/relational aggression (excluding peers from the group, threatening to spread lies or rumors), as well as risk and protective factors and pubertal development. Cross-sectional data were analyzed.

RESULTS Compared with early puberty, the odds of violent behavior were approximately threefold higher in midpuberty (odds ratio [OR]: 2.87 [95% confidence interval (CI): 1.81–4.55]) and late puberty (OR: 3.79 [95% CI: 2.25–6.39]) after adjustment for demographic factors. For social/relational aggression, there were weaker overall associations after adjustment, but these associations included an interaction between pubertal stage and age, and stronger associations with pubertal stage at younger age were shown (P = .003; midpuberty OR: 1.78 [95% CI: 1.20–2.63]; late puberty OR: 3.00 [95% CI: 1.95–4.63]). Associations between pubertal stage and violent behavior and social/relational aggression remained after the inclusion of social contextual mediators in the analyses.

CONCLUSIONS Pubertal stage was associated with higher rates of violent behavior and social/relational aggression, with the latter association seen only at younger ages. Puberty is an important phase at which to implement prevention programs to reduce adolescent violent and antisocial behaviors.

Quality-based SQL: specifying information quality in relational database queries

A Structured Query Language extension uses an estimator module to evaluate quality profiles that rate the accuracy and completeness of query results. Users receive information that matches their defined quality constraints and better serves their data needs.

Profitability performance assessment in the Australian construction industry: a global relational two-stage DEA method

Effective performance measurement drives performance and supports the development of construction. Only minimal literature measuring construction performance, efficiency and effectiveness simultaneously can be identified. A global relational two-stage data envelopment analysis (DEA) method is here proposed in order to produce effective and informative performance results. A relational two-stage DEA method systematically measures overall efficiency for a whole construction system and also yields scores for the individual stages of construction. The DEA results can be directly compared through global benchmark technology. The Australian construction industry is employed in order to implement the new method, in which profitability performance as a vital indicator of business survival, and its two dimensions of efficiency and effectiveness, are measured. The construction profitability performance and efficiency measures obtained provide evidence of underperformance and a slight imbalance in Australia between 1991 and 2012, while the measures obtained for the effectiveness factor indicate better achievement. The approach here developed promotes progress in modelling two-stage performance measurement and it can be replicated worldwide by construction projects, organizations or industries in order to quantify their performance, identify internal inefficiency components and recognize competitive advantages for promoting sustainable development.