Classification tree analysis of postal questionnaire data to identify risk of excessive gestational weight gain

OBJECTIVE: overweight/obese weight status during pregnancy increases risk of a range of adverse health outcomes for mother and child. Whereas identification of those who are overweight/obese pre-pregnancy and in early pregnancy is straightforward, prediction of who will experience excessive gestational weight gain (EGWG), and thus be at greater risk of becoming overweight or obese during pregnancy is more challenging. The present study sought to better identify those at risk of EGWG by exploring pre-pregnancy BMI as well as a range of psychosocial risk factors identified as risk factors in prior research. METHODS: 225 pregnant women completed self-reported via postal survey measures of height, weight, and psychosocial variables at 16-18 weeks gestation, and reported their weight again at 32-34 weeks to calculate GWG. Classification and regression tree analysis (CART) was used to find subgroups in the data with increased risk of EGWG based on their pre-pregnancy BMI and psychosocial risk factor scores at Time 1. FINDINGS: CART confirmed that self-reported BMI status was a strong predictor of EGWG risk for women who were overweight/obese pre-pregnancy. Normal weight women with low motivation to maintain a healthy diet and who reported lower levels of partner support were also at considerable risk of EGWG. IMPLICATIONS FOR PRACTICE: present findings offer support for inclusion of psychosocial measures (in addition to BMI) in early antenatal visits to detect risk of EGWG. However, these findings also underscore the need for further consideration of effect modifiers that place women at increased or decreased risk of EGWG. Proposed additional constructs are discussed to direct further theory-driven research.

Individual variation in the pup attraction call produced by female Australian fur seals during early lactation

Otariid seals (fur seals and sea lions) are colonial breeders with large numbers of females giving birth on land during a synchronous breeding period. Once pups are born, females alternate between feeding their young ashore and foraging at sea. Upon return, both mother and pup must relocate each other and it is thought to be primarily facilitated by vocal recognition. Vocalizations of thirteen female Australian fur seals (Arctocephalus pusillus doriferus) were recorded during the breeding seasons of December 2000 and 2001, when pups are aged from newborns to one month. The pup attraction call was examined to determine whether females produce individually distinct calls which could be used by pups as a basis for vocal recognition. Potential for individual coding, discriminant function analysis (DFA), and classification and regression tree analysis were used to determine which call features were important in separating individuals. Using the results from all three analyses: F0, MIN F and DUR were considered important in separating individuals. In 76% of cases, the PAC was classified to the correct caller, using DFA, suggesting that there is sufficient stereotypy within individual calls, and sufficient variation between them, to enable vocal recognition by pups of this species.

Carpal tunnel syndrome: diagnostic usefulness of sonography

PURPOSE: To prospectively evaluate accuracy of sonography for diagnosis of carpal tunnel syndrome (CTS) in patients clinically suspected of having the disease in one or both hands.
MATERIALS AND METHODS: A prospective cohort of 133 patients suspected of having CTS were referred to a teaching hospital between October 2001 and June 2002 for electrodiagnostic study. One hundred twenty patients (98 women, 22 men; mean age, 49 years; range, 19–83 years) underwent sonography within 1 week after electrodiagnostic study. Radiologist was blinded to electrodiagnostic study results. Seventy-five patients had bilateral symptoms; 23 patients, right-hand symptoms; and 22 patients, left-hand symptoms (total, 195 symptomatic hands). Cross-sectional area of median nerve was measured at three levels: immediately proximal to carpal tunnel inlet, at carpal tunnel inlet, and at carpal tunnel outlet. Flexor retinaculum was used as a landmark to margins of carpal tunnel. Optimal threshold levels (determined with classification and regression tree analysis) for areas proximal to and at tunnel inlet and at tunnel outlet were used to discriminate between patients with and patients without disease. Sensitivity, specificity, and false-positive and false-negative rates were derived on the basis of final diagnosis, which was determined with clinical history and electrodiagnostic study results as reference standard.
RESULTS: For right hands, sonography had sensitivity of 94% (66 of 70); specificity, 65% (17 of 26); false-positive rate, 12% (nine of 75); and false-negative rate, 19% (four of 21) (cutoff, 0.09 cm2 proximal to tunnel inlet and 0.12 cm2 at tunnel outlet). For left hands, sensitivity was 83% (53 of 64); specificity, 73% (24 of 33); false-positive rate, 15% (nine of 62); and false-negative rate, 31% (11 of 35) (cutoff, 0.10 cm2 proximal to tunnel inlet).
CONCLUSION: Sonography is comparable to electrodiagnostic study in diagnosis of CTS and should be considered as initial test of choice for patients suspected of having CTS.

Fault detection and diagnosis of induction motors using motor current signature analysis and a hybrid FMM-CART model

In this paper, a novel approach to detect and classify comprehensive fault conditions of induction motors using a hybrid fuzzy min-max (FMM) neural network and classification and regression tree (CART) is proposed. The hybrid model, known as FMM-CART, exploits the advantages of both FMM and CART for undertaking data classification and rule extraction problems. A series of real experiments is conducted, whereby the motor current signature analysis method is applied to form a database comprising stator current signatures under different motor conditions. The signal harmonics from the power spectral density are extracted as discriminative input features for fault detection and classification with FMM-CART. A comprehensive list of induction motor fault conditions, viz., broken rotor bars, unbalanced voltages, stator winding faults, and eccentricity problems, has been successfully classified using FMM-CART with good accuracy rates. The results are comparable, if not better, than those reported in the literature. Useful explanatory rules in the form of a decision tree are also elicited from FMM-CART to analyze and understand different fault conditions of induction motors.

Analysis of the stages of change model of drug and alcohol treatment readiness among prisoners

A classification and regression tree (CART) analysis was applied to data for 237 male participants (M = 31.93 years, SD = 7.64) in a prison-based substance abuse treatment program to study the integrity of the Stages of Change model of treatment readiness. Using the Stages of Change Questionnaire (STOCQ), participants were assigned to Contemplation (102), Action (118), or Maintenance (17) groups. A CART analysis then examined differences in the overall group profiles on the basis of scores on the Psychological Inventory of Criminal Thinking, the Situational Confidence Questionnaire, and the Carlson Psychological Survey. The assumption of discrete stages of change was not supported. Alternative models are suggested: one based on states of change and one on personality characteristics. A focus on equal attention to both cognitive and behavioral aspects of substance abuse treatment readiness is suggested.

A hybrid FAM–CART model and its application to medical data classification

In this paper, a hybrid model consisting of the fuzzy ARTMAP (FAM) neural network and the classification and regression tree (CART) is formulated. FAM is useful for tackling the stability–plasticity dilemma pertaining to data-based learning systems, while CART is useful for depicting its learned knowledge explicitly in a tree structure. By combining the benefits of both models, FAM–CART is capable of learning data samples stably and, at the same time, explaining its predictions with a set of decision rules. In other words, FAM–CART possesses two important properties of an intelligent system, i.e., learning in a stable manner (by overcoming the stability–plasticity dilemma) and extracting useful explanatory rules (by overcoming the opaqueness issue). To evaluate the usefulness of FAM–CART, six benchmark medical data sets from the UCI repository of machine learning and a real-world medical data classification problem are used for evaluation. For performance comparison, a number of performance metrics which include accuracy, specificity, sensitivity, and the area under the receiver operation characteristic curve are computed. The results are quantified with statistical indicators and compared with those reported in the literature. The outcomes positively indicate that FAM–CART is effective for undertaking data classification tasks. In addition to producing good results, it provides justifications of the predictions in the form of a decision tree so that domain users can easily understand the predictions, therefore making it a useful decision support tool.

Whale distribution in relation to prey abundance and oceanographic processes in the shelf waters of the Western Antarctic Peninsula

The Western Antarctic Peninsula (WAP) is a biologically rich area supporting large standing stocks of krill and top predators (including whales, seals and seabirds). Physical forcing greatly affects productivity, recruitment, survival and distribution of krill in this area. In turn, such interactions are likely to affect the distribution of baleen whales. The Southern Ocean GLOBEC research program aims to explore the relationships and interactions between the environment, krill and predators around Marguerite Bay (WAP) in autumn 2001 and 2002. Bathymetric and environmental variables including acoustic backscattering as an indicator of prey abundance were used to model whale distribution patterns. We used an iterative approach employing (1) classification and regression tree (CART) models to identify oceanographic and ecological variables contributing to variability in humpback Megaptera novaeangliae and minke Balaenoptera acutorstrata whale distribution, and (2) generalized additive models (GAMs) to elucidate functional ecological relationships between these variables and whale distribution. The CART models indicated that the cetacean distribution was tightly coupled with zooplankton acoustic volume backscatter in the upper (25 to 100 m), and middle (100 to 300 m) portions of the water column. Whale distribution was also related to distance from the ice edge and bathymetric slope. The GAMs indicated a persistent, strong, positive relationship between increasing zooplankton volume and whale relative abundance. Furthermore, there was a lower limit for averaged acoustic volume backscatter of zooplankton below which the relationship between whales and prey was not significant. The GAMs also supported an annual relationship between whale distribution, distance from the ice edge and bathymetric slope, suggesting that these are important features for aggregating prey. Our results demonstrate that during the 2 yr study, whales were consistently and predictably associated with the distribution of zooplankton. Thus, humpback and minke whales may be able to locate physical features and oceanographic processes that enhance prey aggregation.

Lymphedema after complete axillary node dissection for melanoma: assessment using a new, objective definition

Objectives: The objectives of this study were to define appropriate criteria for assessing the presence of lymphedema, and to report the prevalence and risk factors for development of upper limb lymphedema after level I-III axillary dissection for melanoma.
Summary Background Data: The lack of a consistent and reliable objective definition for lymphedema remains a significant barrier to appreciating its prevalence after axillary dissection for melanoma (or breast carcinoma).
Methods: Lymphedema was assessed in 107 patients (82 male, 25 female) who had previously undergone complete level I-III axillary dissection. Of the 107 patients, 17 had also received postoperative axillary radiotherapy. Arm volume was measured using a water displacement technique. Change in volume of the arm on the side of the dissection was referenced to the volume of the other (control) arm. Volume measurements were corrected for the effect of handedness using corrections derived from a control group. Classification and regression tree (CART) analysis was used to determine a threshold fractional arm volume increase above which volume changes were considered to indicate lymphedema.
Results: Based on the CART analysis results, lymphedema was defined as an increase in arm volume greater than 16% of the volume of the control arm. Using this definition, lymphedema prevalence for patients in the present study was 10% after complete level I-III axillary dissection for melanoma and 53% after additional axillary radiotherapy. Radiotherapy and wound complications were independent risk factors for the development of lymphedema.
Conclusions: A suggested objective definition for arm lymphedema after axillary dissection is an arm volume increase of greater than 16% of the volume of the control arm.

Predicting future ecological degradation based on modelled thresholds

Threshold models are becoming important in determining the ecological consequences of our actions within the environment and have a key role in setting bounds on targets used by natural resource managers. We have been using thresholds and related concepts adapted from the multiple stable-states literature to model ecosystem response in the Coorong, the estuary for Australia’s largest river. Our modelling approach is based upon developing a state-and-transition model, with the states defined by the biota and the transitions defined by a classification and regression tree (CART) analysis of the environmental data for the region. Here we explore the behaviour of thresholds within that model. Managers tend to plan for a set of often arbitrarily-derived thresholds in their natural resource management. We attempt to assess how the precision afforded by analyses such as CART translates into ecological outcomes, and explicitly trial several approaches to understanding thresholds and transitions in our model and how they might be relevant for management. We conclude that the most promising approach would be a mixture of further modelling (using past behaviour to predict future degradation) in conjunction with targeted experiments to confirm the results. Our case study of the Coorong is further developed, particularly for the modelling stages of the protocol, to provide recommendations to improve natural resource management strategies that are currently in use.

Online motor fault detection and diagnosis using a hybrid FMM-CART model

In this brief, a hybrid model combining the fuzzy min-max (FMM) neural network and the classification and regression tree (CART) for online motor detection and diagnosis tasks is described. The hybrid model, known as FMM-CART, exploits the advantages of both FMM and CART for undertaking data classification and rule extraction problems. To evaluate the applicability of the proposed FMM-CART model, an evaluation with a benchmark data set pertaining to electrical motor bearing faults is first conducted. The results obtained are equivalent to those reported in the literature. Then, a laboratory experiment for detecting and diagnosing eccentricity faults in an induction motor is performed. In addition to producing accurate results, useful rules in the form of a decision tree are extracted to provide explanation and justification for the predictions from FMM-CART. The experimental outcome positively shows the potential of FMM-CART in undertaking online motor fault detection and diagnosis tasks.