Final Report : learning system for life prediction of infrastructure

The project has further developed two programs for the industry partners related to service life prediction and salt deposition. The program for Queensland Department of Main Roads which predicts salt deposition on different bridge structures at any point in Queensland has been further refined by looking at more variables. It was found that the height of the bridge significantly affects the salt deposition levels only when very close to the coast. However the effect of natural cleaning of salt by rainfall was incorporated into the program. The user interface allows selection of a location in Queensland, followed by a bridge component. The program then predicts the annual salt deposition rate and rates the likely severity of the environment. The service life prediction program for the Queensland Department of Public Works has been expanded to include 10 common building components, in a variety of environments. Data mining procedures have been used to develop the program and increase the usefulness of the application. A Query Based Learning System (QBLS) has been developed which is based on a data-centric model with extensions to provide support for user interaction. The program is based on number of sources of information about the service life of building components. These include the Delphi survey, the CSIRO Holistic model and a school survey. During the project, the Holistic model was modified for each building component and databases generated for the locations of all Queensland schools. Experiments were carried out to verify and provide parameters for the modelling. These included instrumentation of a downpipe, measurements on pH and chloride levels in leaf litter, EIS measurements and chromate leaching from Colorbond materials and dose tests to measure corrosion rates of new materials. A further database was also generated for inclusion in the program through a large school survey. Over 30 schools in a range of environments from tropical coastal to temperate inland were visited and the condition of the building components rated on a scale of 0-5. The data was analysed and used to calculate an average service life for each component/material combination in the environments, where sufficient examples were available.

Report : stochastic analysis of road asset data for maintenance cost prediction

Reliable budget/cost estimates for road maintenance and rehabilitation are subjected to uncertainties and variability in road asset condition and characteristics of road users. The CRC CI research project 2003-029-C ‘Maintenance Cost Prediction for Road’ developed a method for assessing variation and reliability in budget/cost estimates for road maintenance and rehabilitation. The method is based on probability-based reliable theory and statistical method. The next stage of the current project is to apply the developed method to predict maintenance/rehabilitation budgets/costs of large networks for strategic investment. The first task is to assess the variability of road data. This report presents initial results of the analysis in assessing the variability of road data. A case study of the analysis for dry non reactive soil is presented to demonstrate the concept in analysing the variability of road data for large road networks. In assessing the variability of road data, large road networks were categorised into categories with common characteristics according to soil and climatic conditions, pavement conditions, pavement types, surface types and annual average daily traffic. The probability distributions, statistical means, and standard deviation values of asset conditions and annual average daily traffic for each type were quantified. The probability distributions and the statistical information obtained in this analysis will be used to asset the variation and reliability in budget/cost estimates in later stage. Generally, we usually used mean values of asset data of each category as input values for investment analysis. The variability of asset data in each category is not taken into account. This analysis method demonstrated that it can be used for practical application taking into account the variability of road data in analysing large road networks for maintenance/rehabilitation investment analysis.

The user query based learning system for lifetime prediction of metallic components

Real-World Data Mining Applications generally do not end up with the creation of the models. The use of the model is the final purpose especially in prediction tasks. The problem arises when the model is built based on much more information than that the user can provide in using the model. As a result, the performance of model reduces drastically due to many missing attributes values. This paper develops a new learning system framework, called as User Query Based Learning System (UQBLS), for building data mining models best suitable for users use. We demonstrate its deployment in a real-world application of the lifetime prediction of metallic components in buildings

Bayesian model averaging for harmful algal bloom prediction

Harmful Algal Blooms (HABs) are a worldwide problem that have been increasing in frequency and extent over the past several decades. HABs severely damage aquatic ecosystems by destroying benthic habitat, reducing invertebrate and fish populations and affecting larger species such as dugong that rely on seagrasses for food. Few statistical models for predicting HAB occurrences have been developed, and in common with most predictive models in ecology, those that have been developed do not fully account for uncertainties in parameters and model structure. This makes management decisions based on these predictions more risky than might be supposed. We used a probit time series model and Bayesian Model Averaging (BMA) to predict occurrences of blooms of Lyngbya majuscula, a toxic cyanophyte, in Deception Bay, Queensland, Australia. We found a suite of useful predictors for HAB occurrence, with Temperature figuring prominently in models with the majority of posterior support, and a model consisting of the single covariate average monthly minimum temperature showed by far the greatest posterior support. A comparison of alternative model averaging strategies was made with one strategy using the full posterior distribution and a simpler approach that utilised the majority of the posterior distribution for predictions but with vastly fewer models. Both BMA approaches showed excellent predictive performance with little difference in their predictive capacity. Applications of BMA are still rare in ecology, particularly in management settings. This study demonstrates the power of BMA as an important management tool that is capable of high predictive performance while fully accounting for both parameter and model uncertainty.

Prediction of highway truck conveyance configuration under over-limit prohibition and conveyance equilibrium

Aiming at the shortage of prevailing prediction methods about highway truck conveyance configuration in over-limit freight research that transferring the goods attributed to over-limit portion to another fully loaded truck of the same configuration and developing the truck traffic volume synchronously, a new way to get accumulated probability function of truck power tonnage in basal year by highway truck classified by wheel and axle type load mass spectrum investigation was presented. Logit models were used to forecast overall highway freight diversion and single cargo tonnage diversion when the weight rules and strict of enforcement intensity of overload were changed in scheme year. Assumption that the probability distribution of single truck loadage should be consistent with the probability distribution of single goods freighted, the model describes the truck conveyance configuration in the future under strict over-limit prohibition. The model was used and tested in Highway Over-limit Research Project in Anhui by World Bank.

Patterns of query reformulation during web searching

Query reformulation is a key user behavior during Web search. Our research goal is to develop predictive models of query reformulation during Web searching. This article reports results from a study in which we automatically classified the query-reformulation patterns for 964,780 Web searching sessions, composed of 1,523,072 queries, to predict the next query reformulation. We employed an n-gram modeling approach to describe the probability of users transitioning from one query-reformulation state to another to predict their next state. We developed first-, second-, third-, and fourth-order models and evaluated each model for accuracy of prediction, coverage of the dataset, and complexity of the possible pattern set. The results show that Reformulation and Assistance account for approximately 45% of all query reformulations; furthermore, the results demonstrate that the first- and second-order models provide the best predictability, between 28 and 40% overall and higher than 70% for some patterns. Implications are that the n-gram approach can be used for improving searching systems and searching assistance.

Predicting query reformulation during web searching

This paper reports results from a study in which we automatically classified the query reformulation patterns for 964,780 Web searching sessions (composed of 1,523,072 queries) in order to predict what the next query reformulation would be. We employed an n-gram modeling approach to describe the probability of searchers transitioning from one query reformulation state to another and predict their next state. We developed first, second, third, and fourth order models and evaluated each model for accuracy of prediction. Findings show that Reformulation and Assistance account for approximately 45 percent of all query reformulations. Searchers seem to seek system searching assistant early in the session or after a content change. The results of our evaluations show that the first and second order models provided the best predictability, between 28 and 40 percent overall, and higher than 70 percent for some patterns. Implications are that the n-gram approach can be used for improving searching systems and searching assistance in real time.

Prediction of adherence and control in diabetes

This study aims to predict adherence to diabetic treatment regimens and sustained diabetic control. During two clinic visits that were 2 months apart, 63 adult outpatients completed measures of diabetic history, current treatment, diabetic control, adherence, and self-efficacy about adherence to treatment. Results showed that self-efficacy was a significant predictor of later adherence to diabetes treatment even after past levels of adherence were taken into account. Posttest levels of adherence in turn were significantly associated with posttest %HbA1c after control for illness severity. A stepwise multiple regression to predict %HbAlc at post entered pretest measures of diabetic control, treatment type, and self-efficacy, which together predicted 50% of the variance. Results are related to self-efficacy theory and implications for practice are discussed.

The Alcohol Helplessness Scale and its prediction of depression among problem drinkers

Event-specific scales commonly have greater power than generalized scales in prediction of specific disorders and in testing mediator models for predicting such disorders. Therefore, in a preliminary study, a 6-item Alcohol Helplessness Scale was constructed and found to be reliable for a sample of 98 problem drinkers. Hierarchical multiple regression and its derivative path analysis were used to test whether helplessness and self-efficacy moderate or mediate the link between alcohol dependence and depression, A test of a moderation model was not supported, whereas a test of a mediation model was supported. Helplessness and self-efficacy both significantly and independently mediated between alcohol dependence and depression. Nevertheless, a significant direct effect of alcohol dependence on depression also remained.

Prediction of psychological adjustment to multiple sclerosis

This study examined the utility of self-efficacy as a predictor of social activity and mood control in multiple sclerosis (MS). Seventy-one subjects with MS were recruited from people attending an MS centre or from a mailing list and were examined on two occasions that were two months apart. Clinic patients were more disabled than patients who completed assessments by post, but they were of higher socioeconomic status and were less dysphoric. We attempted to predict self-reported performance of mood control and social activity at two months, from self-efficacy or performance on these tasks at pretest. Demographic variables, disorder status, disability, self-esteem and depression were also allowed to compete for entry into multiple regressions. Substantial stability in mood, performance and disability was observed over the two months. In both mood control and social activity, past performance was the strongest predictor of later performance, but self-efficacy also contributed significantly to the prediction. The disability level entered a prediction of socila activity, but no other variables predicted either type of performance. A secondary analysis predicting self-esteem at two months also included self-efficacy for social activity, illustrating the contribution of perceived capability to later assessments of self-worth. The study provided support for self-efficacy as a predictor of later behavioural outcomes and self-esteem in multiple sclerosis.

Prediction of adjacent vertebral fracture risk associated with vertebroplasty : a computer-based simulation study

Vertebrplasty involved injecting cement into a fractured vertebra to provide stabilisation. There is clinical evidence to suggest however that vertebroplasty may be assocated with a higher risk of adjacent vertebral fracture; which may be due to the change in material properties of the post-procedure vertebra modifying the transmission of mechanical stresses to adjacent vertebrae.

Cough-generated aerosols of Pseudomonas aeruginosa and other Gram Negative Bacteria from cystic fibrosis patients.

Background: Pseudomonas aeruginosa is the most common bacterial pathogen in cystic fibrosis (CF) patients. Current infection control guidelines aim to prevent transmission via contact and respiratory droplet routes and do not consider the possibility of airborne transmission. We hypothesized that with coughing, CF subjects produce viable, respirable bacterial aerosols. Methods: Cross-sectional study of 15 children and 13 adults with CF, 26 chronically infected with P. aeruginosa. A cough aerosol sampling system enabled fractioning of respiratory particles of different size, and culture of viable Gram negative non-fermentative bacteria. We collected cough aerosols during 5 minutes voluntary coughing and during a sputum induction procedure when tolerated. Standardized quantitative culture and genotyping techniques were used. Results: P. aeruginosa was isolated in cough aerosols of 25 (89%) subjects of whom 22 produced sputum samples. P. aeruginosa from sputum and paired cough aerosols were indistinguishable by molecular typing. In 4 cases the same genotype was isolated from ambient room air. Approximately 70% of viable aerosols collected during voluntary coughing were of particles ≤ 3.3 microns aerodynamic diameter. P. aeruginosa, Burkholderia cenocepacia Stenotrophomonas maltophilia and Achromobacter xylosoxidans were cultivated from respiratory particles in this size range. Positive room air samples were associated with high total counts in cough aerosols (P=0.003). The magnitude of cough aerosols were associated with higher FEV1 (r=0.45, P=0.02) and higher quantitative sputum culture results (r=0.58, P=0.008). Conclusion: During coughing, CF patients produce viable aerosols of P. aeruginosa and other Gram negative bacteria of respirable size range, suggesting the potential for airborne transmission.

Prediction of outcome with group cognitive therapy for depression

Tested a social–cognitive model of depressive episodes and their treatment within a predictive study of treatment response. 42 clinically depressed volunteers (aged 22–60 yrs) were given self-efficacy (SE) questionnaires and other measures before and after treatment with cognitive therapy. Results support the idea that SE and skills regarding control of negative cognition mediates a sustained response to cognitive treatment for depression. Not only did mood-control variables correlate highly with concurrent changes in depression scores during treatment, but the posttreatment SE measure discriminated Ss who relapsed over the next 12 mo.

Word Segmentation for Chinese Wikipedia Using N-Gram Mutual Information

In this paper, we propose an unsupervised segmentation approach, named "n-gram mutual information", or NGMI, which is used to segment Chinese documents into n-character words or phrases, using language statistics drawn from the Chinese Wikipedia corpus. The approach alleviates the tremendous effort that is required in preparing and maintaining the manually segmented Chinese text for training purposes, and manually maintaining ever expanding lexicons. Previously, mutual information was used to achieve automated segmentation into 2-character words. The NGMI approach extends the approach to handle longer n-character words. Experiments with heterogeneous documents from the Chinese Wikipedia collection show good results.