Minimizing the effects of overfitting and collinearity in construction cost estimation : a new hybird approach

Research problem: Overfitting and collinearity problems commonly exist in current construction cost estimation applications and obstruct researchers and practitioners in achieving better modelling results. Research objective and method: A hybrid approach of Akaike information criterion (AIC) stepwise regression and principal component regression (PCR) is proposed to help solve overfitting and collinearity problems. Utilization of this approach in linear regression is validated by comparing it with other commonly used approaches. The mean square error obtained by leave-one-out cross validation (MSELOOCV) is used in model selection in deciding predictive variables.

An accuracy assessment of ASTER stereo images-derived digital elevation model by using rational polynomial coefficient model

The along-track stereo images of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) sensor with 15 m resolution were used to generate Digital Elevation Model (DEM) on an area with low and near Mean Sea Level (MSL) elevation in Johor, Malaysia. The absolute DEM was generated by using the Rational Polynomial Coefficient (RPC) model which was run on ENVI 4.8 software. In order to generate the absolute DEM, 60 Ground Control Pointes (GCPs) with almost vertical accuracy less than 10 meter extracted from topographic map of the study area. The assessment was carried out on uncorrected and corrected DEM by utilizing dozens of Independent Check Points (ICPs). Consequently, the uncorrected DEM showed the RMSEz of ± 26.43 meter which was decreased to the RMSEz of ± 16.49 meter for the corrected DEM after post-processing. Overall, the corrected DEM of ASTER stereo images met the expectations.

Population variation in biomonitoring data for persistent organic pollutants (POPs) : an examination of multiple population-based datasets for application to Australian pooled biomonitoring data

Background Australian national biomonitoring for persistent organic pollutants (POPs) relies upon age-specific pooled serum samples to characterize central tendencies of concentrations but does not provide estimates of upper bound concentrations. This analysis compares population variation from biomonitoring datasets from the US, Canada, Germany, Spain, and Belgium to identify and test patterns potentially useful for estimating population upper bound reference values for the Australian population. Methods Arithmetic means and the ratio of the 95th percentile to the arithmetic mean (P95:mean) were assessed by survey for defined age subgroups for three polychlorinated biphenyls (PCBs 138, 153, and 180), hexachlorobenzene (HCB), p,p-dichlorodiphenyldichloroethylene (DDE), 2,2′,4,4′ tetrabrominated diphenylether (PBDE 47), perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS). Results Arithmetic mean concentrations of each analyte varied widely across surveys and age groups. However, P95:mean ratios differed to a limited extent, with no systematic variation across ages. The average P95:mean ratios were 2.2 for the three PCBs and HCB; 3.0 for DDE; 2.0 and 2.3 for PFOA and PFOS, respectively. The P95:mean ratio for PBDE 47 was more variable among age groups, ranging from 2.7 to 4.8. The average P95:mean ratios accurately estimated age group-specific P95s in the Flemish Environmental Health Survey II and were used to estimate the P95s for the Australian population by age group from the pooled biomonitoring data. Conclusions Similar population variation patterns for POPs were observed across multiple surveys, even when absolute concentrations differed widely. These patterns can be used to estimate population upper bounds when only pooled sampling data are available.

Deriving a preference-based utility measure for cancer patients from the European Organisation for the Research and Treatment of Cancer's Quality of Life Questionnaire C30 : a confirmatory versus exploratory approach

Background Multi attribute utility instruments (MAUIs) are preference-based measures that comprise a health state classification system (HSCS) and a scoring algorithm that assigns a utility value to each health state in the HSCS. When developing a MAUI from a health-related quality of life (HRQOL) questionnaire, first a HSCS must be derived. This typically involves selecting a subset of domains and items because HRQOL questionnaires typically have too many items to be amendable to the valuation task required to develop the scoring algorithm for a MAUI. Currently, exploratory factor analysis (EFA) followed by Rasch analysis is recommended for deriving a MAUI from a HRQOL measure. Aim To determine whether confirmatory factor analysis (CFA) is more appropriate and efficient than EFA to derive a HSCS from the European Organisation for the Research and Treatment of Cancer’s core HRQOL questionnaire, Quality of Life Questionnaire (QLQ-C30), given its well-established domain structure. Methods QLQ-C30 (Version 3) data were collected from 356 patients receiving palliative radiotherapy for recurrent/metastatic cancer (various primary sites). The dimensional structure of the QLQ-C30 was tested with EFA and CFA, the latter informed by the established QLQ-C30 structure and views of both patients and clinicians on which are the most relevant items. Dimensions determined by EFA or CFA were then subjected to Rasch analysis. Results CFA results generally supported the proposed QLQ-C30 structure (comparative fit index =0.99, Tucker–Lewis index =0.99, root mean square error of approximation =0.04). EFA revealed fewer factors and some items cross-loaded on multiple factors. Further assessment of dimensionality with Rasch analysis allowed better alignment of the EFA dimensions with those detected by CFA. Conclusion CFA was more appropriate and efficient than EFA in producing clinically interpretable results for the HSCS for a proposed new cancer-specific MAUI. Our findings suggest that CFA should be recommended generally when deriving a preference-based measure from a HRQOL measure that has an established domain structure.

Simulated hyperopic anisometropia and reading, visual information processing, and reading-related eye movement performance in children

Purpose: This study investigated the impact of simulated hyperopic anisometropia and sustained near work on performance of academic-related measures in children. Methods: Participants included 16 children (mean age: 11.1 ± 0.8 years) with minimal refractive error. Academic-related outcome measures included a reading test (Neale Analysis of Reading Ability), visual information processing tests (Coding and Symbol Search subtests from the Wechsler Intelligence Scale for Children) and a reading-related eye movement test (Developmental Eye Movement test). Performance was assessed with and without 0.75 D of imposed monocular hyperopic defocus (administered in a randomised order), before and after 20 minutes of sustained near work. Unilateral hyperopic defocus was systematically assigned to either the dominant or non-dominant sighting eye to evaluate the impact of ocular dominance on any performance decrements. Results: Simulated hyperopic anisometropia and sustained near work both independently reduced performance on all of the outcome measures (p<0.001). A significant interaction was also observed between simulated anisometropia and near work (p<0.05), with the greatest decrement in performance observed during simulated anisometropia in combination with sustained near work. Laterality of the refractive error simulation (ocular dominance) did not significantly influence the outcome measures (p>0.05). A reduction of up to 12% in performance was observed across the range of academic-related measures following sustained near work undertaken during the anisometropic simulation. Conclusion: Simulated hyperopic anisometropia significantly impaired academic–related performance, particularly in combination with sustained near work. The impact of uncorrected habitual anisometropia on academic-related performance in children requires further investigation.

Cone Ratios in Myopia and Emmetropia: A Pilot Study

Purpose There is a suggestion that the long wavelength-sensitive (LWS)-to-middle wavelength-sensitive (MWS) cone ratio in the retina is associated with myopia. The aim was to measure the LWS/MWS amplitude modulation ratio, an estimate of the LWS/MWS cone ratio, in young adult emmetropes and myopes. Methods Multifocal visual evoked potentials were measured when the LWS and MWS cone systems were excited separately using the method of silent substitution. The 30 young adult participants (22 to 33 years) included 10 emmetropes (mean [±SD] refraction, +0.3 [±0.4] diopters [D]) and 20 myopes (mean [±SD] refraction, -3.4 [±1.7] D). Results The LWS/MWS amplitude modulation ratios ranged from 0.56 to 1.80 in the central 3- to 13-degree diameter ring and from 0.94 to 1.91 in the peripheral 13- to 30-degree diameter ring. Within the central ring, the mean (±SD) ratios were 1.20 (±0.26) and 1.20 (±0.33) for the emmetropic and the myopic groups, respectively. For the peripheral ring, the mean (±SD) ratios were 1.48 (±0.27) and 1.30 (±0.27), respectively. There were no significant differences in the ratios between the emmetropic and myopic groups for either the central (p = 0.99) or peripheral (p = 0.08) rings. For the latter, more myopic refractive error was associated with lower LWS/MWS amplitude modulation ratio; the refraction explained 16% (p = 0.02) of variation in ratio. Conclusions The relationship between the LWS/MWS amplitude modulation ratios and refraction at 13 to 30 degrees indicates that a large longitudinal study of changes in refraction in persons with known cone ratio is required to determine if a low LWS/MWS cone ratio is associated with myopia development.

Differential effects of refractive blur on day and night-time driving performance

Purpose To investigate the effect of different levels of refractive blur on real-world driving performance measured under day and nighttime conditions. Methods Participants included 12 visually normal, young adults (mean age = 25.8 ± 5.2 years) who drove an instrumented research vehicle around a 4 km closed road circuit with three different levels of binocular spherical refractive blur (+0.50 diopter sphere [DS], +1.00 DS, +2.00 DS) compared with a baseline condition. The subjects wore optimal spherocylinder correction and the additional blur lenses were mounted in modified full-field goggles; the order of testing of the blur conditions was randomized. Driving performance was assessed in two different sessions under day and nighttime conditions and included measures of road signs recognized, hazard detection and avoidance, gap detection, lane-keeping, sign recognition distance, speed, and time to complete the course. Results Refractive blur and time of day had significant effects on driving performance (P < 0.05), where increasing blur and nighttime driving reduced performance on all driving tasks except gap judgment and lane keeping. There was also a significant interaction between blur and time of day (P < 0.05), such that the effects of blur were exacerbated under nighttime driving conditions; performance differences were evident even for +0.50 DS blur relative to baseline for some measures. Conclusions The effects of blur were greatest under nighttime conditions, even for levels of binocular refractive blur as low as +0.50 DS. These results emphasize the importance of accurate and up-to-date refractive correction of even low levels of refractive error when driving at night.

Combining opinions for use in Bayesian networks: A measurement error approach

Bayesian networks (BNs) are graphical probabilistic models used for reasoning under uncertainty. These models are becoming increasing popular in a range of fields including ecology, computational biology, medical diagnosis, and forensics. In most of these cases, the BNs are quantified using information from experts, or from user opinions. An interest therefore lies in the way in which multiple opinions can be represented and used in a BN. This paper proposes the use of a measurement error model to combine opinions for use in the quantification of a BN. The multiple opinions are treated as a realisation of measurement error and the model uses the posterior probabilities ascribed to each node in the BN which are computed from the prior information given by each expert. The proposed model addresses the issues associated with current methods of combining opinions such as the absence of a coherent probability model, the lack of the conditional independence structure of the BN being maintained, and the provision of only a point estimate for the consensus. The proposed model is applied an existing Bayesian Network and performed well when compared to existing methods of combining opinions.

The Craving Experience Questionnaire : a brief, theory-based measure of consummatory desire and craving

Background and Aims Research into craving is hampered by lack of theoretical specification and a plethora of substance-specific measures. This study aimed to develop a generic measure of craving based on elaborated intrusion (EI) theory. Confirmatory factor analysis (CFA) examined whether a generic measure replicated the three-factor structure of the Alcohol Craving Experience (ACE) scale over different consummatory targets and time-frames. Design Twelve studies were pooled for CFA. Targets included alcohol, cigarettes, chocolate and food. Focal periods varied from the present moment to the previous week. Separate analyses were conducted for strength and frequency forms. Setting Nine studies included university students, with single studies drawn from an internet survey, a community sample of smokers and alcohol-dependent out-patients. Participants A heterogeneous sample of 1230 participants. Measurements Adaptations of the ACE questionnaire. Findings Both craving strength [comparative fit indices (CFI = 0.974; root mean square error of approximation (RMSEA) = 0.039, 95% confidence interval (CI) = 0.035–0.044] and frequency (CFI = 0.971, RMSEA = 0.049, 95% CI = 0.044–0.055) gave an acceptable three-factor solution across desired targets that mapped onto the structure of the original ACE (intensity, imagery, intrusiveness), after removing an item, re-allocating another and taking intercorrelated error terms into account. Similar structures were obtained across time-frames and targets. Preliminary validity data on the resulting 10-item Craving Experience Questionnaire (CEQ) for cigarettes and alcohol were strong. Conclusions The Craving Experience Questionnaire (CEQ) is a brief, conceptually grounded and psychometrically sound measure of desires. It demonstrates a consistent factor structure across a range of consummatory targets in both laboratory and clinical contexts.

Causal factors of hot air ballooning incidents : identification, frequency, and potential impact

Background: Hot air ballooning incidents are relatively rare, however, when they do occur they are likely to result in a fatality or serious injury. Human error is commonly attributed as the cause of hot air ballooning incidents; however, error in itself is not an explanation for safety failures. This research aims to identify, and establish the relative importance of factors contributing towards hot air ballooning incidents. Methods: Twenty-two Australian Ballooning Federation (ABF) incident reports were thematically coded using a bottom up approach to identify causal factors. Subsequently, 69 balloonists (mean 19.51 years’ experience) participated in a survey to identify additional causal factors and rate (out of seven) the perceived frequency and potential impact to ballooning operations of each of the previously identified causal factors. Perceived associated risk was calculated by multiplying mean perceived frequency and impact ratings. Results: Incident report coding identified 54 causal factors within nine higher level areas: Attributes, Crew resource management, Equipment, Errors, Instructors, Organisational, Physical Environment, Regulatory body and Violations. Overall, ‘weather’, ‘inexperience’ and ‘poor/inappropriate decisions’ were rated as having greatest perceived associated risk. Discussion: Although errors were nominated as a prominent cause of hot air ballooning incidents, physical environment and personal attributes are also particularly important for safe hot air ballooning operations. In identifying a range of causal factors the areas of weakness surrounding ballooning operations have been defined; it is hoped that targeted safety and training strategies can now be put into place removing these contributing factors and reducing the chance of pilot error.

Influences of luminance and accommodation stimuli on pupil size and pupil center location

PURPOSE. To investigate effects of luminance and accommodation stimuli on pupil size and pupil center location and their implications for progressive addition lens wear. METHODS. Participants were young and older adult groups (n=20, 22±2 years, age range 18-25 years; n=19, 49±4 years, 45-58 years). A wave aberrometer included a relay system to allow a 12.5°x11° background for the internal fixation target. Participants viewed the target under a matrix of conditions with luminance levels 0.01, 3.7, 120 and 6100 cd/m² and with accommodation stimuli up to 6 diopters in 2 diopter steps. Pupil sizes and their centers, relative to limbus centers, were determined from anterior eye images. RESULTS. With luminance increase, reduction in pupil size was accentuated by increase in accommodation stimulus in the young, but not in the older, group. As luminance increased, pupil center location altered. This was nasally in both groups with an average shift of approximately 0.12mm. Relative to the lowest stimulus condition, the mean of the maximum absolute pupil center shifts was 0.26±0.08mm for both groups with individual shifts up to 0.5mm, findings consistent with previous studies. There was no significant effect of accommodation on pupil center locations for either age group, or evidence that location was influenced by the combination of luminance and accommodation stimulus that resulted in any particular pupil size. CONCLUSIONS. Variations in luminance and accommodation influence pupil size, but only the former affects pupil center location significantly. Pupil center shifts are too small to be of concern in fitting progressive addition lenses.

Acoustic identification of 12 species of echolocating bat by discriminant function analysis and artificial neural networks

We recorded echolocation calls from 14 sympatric species of bat in Britain. Once digitised, one temporal and four spectral features were measured from each call. The frequency-time course of each call was approximated by fitting eight mathematical functions, and the goodness of fit, represented by the mean-squared error, was calculated. Measurements were taken using an automated process that extracted a single call from background noise and measured all variables without intervention. Two species of Rhinolophus were easily identified from call duration and spectral measurements. For the remaining 12 species, discriminant function analysis and multilayer back-propagation perceptrons were used to classify calls to species level. Analyses were carried out with and without the inclusion of curve-fitting data to evaluate its usefulness in distinguishing among species. Discriminant function analysis achieved an overall correct classification rate of 79% with curve-fitting data included, while an artificial neural network achieved 87%. The removal of curve-fitting data improved the performance of the discriminant function analysis by 2 %, while the performance of a perceptron decreased by 2 %. However, an increase in correct identification rates when curve-fitting information was included was not found for all species. The use of a hierarchical classification system, whereby calls were first classified to genus level and then to species level, had little effect on correct classification rates by discriminant function analysis but did improve rates achieved by perceptrons. This is the first published study to use artificial neural networks to classify the echolocation calls of bats to species level. Our findings are discussed in terms of recent advances in recording and analysis technologies, and are related to factors causing convergence and divergence of echolocation call design in bats.

Error and variance bounds on sigmoidal neurons with weight and input errors

Bounds on the expectation and variance of errors at the output of a multilayer feedforward neural network with perturbed weights and inputs are derived. It is assumed that errors in weights and inputs to the network are statistically independent and small. The bounds obtained are applicable to both digital and analogue network implementations and are shown to be of practical value.