Predictive models for chronic renal disease using decision trees, naïve bayes and case-based methods

Data mining can be used in healthcare industry to “mine” clinical data to discover hidden information for intelligent and affective decision making. Discovery of hidden patterns and relationships often goes intact, yet advanced data mining techniques can be helpful as remedy to this scenario. This thesis mainly deals with Intelligent Prediction of Chronic Renal Disease (IPCRD). Data covers blood, urine test, and external symptoms applied to predict chronic renal disease. Data from the database is initially transformed to Weka (3.6) and Chi-Square method is used for features section. After normalizing data, three classifiers were applied and efficiency of output is evaluated. Mainly, three classifiers are analyzed: Decision Tree, Naïve Bayes, K-Nearest Neighbour algorithm. Results show that each technique has its unique strength in realizing the objectives of the defined mining goals. Efficiency of Decision Tree and KNN was almost same but Naïve Bayes proved a comparative edge over others. Further sensitivity and specificity tests are used as statistical measures to examine the performance of a binary classification. Sensitivity (also called recall rate in some fields) measures the proportion of actual positives which are correctly identified while Specificity measures the proportion of negatives which are correctly identified. CRISP-DM methodology is applied to build the mining models. It consists of six major phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment.

Parameter estimation of Poisson generalized linear mixed models based on three different statistical principles : a simulation study

Generalized linear mixed models are flexible tools for modeling non-normal data and are useful for accommodating overdispersion in Poisson regression models with random effects. Their main difficulty resides in the parameter estimation because there is no analytic solution for the maximization of the marginal likelihood. Many methods have been proposed for this purpose and many of them are implemented in software packages. The purpose of this study is to compare the performance of three different statistical principles - marginal likelihood, extended likelihood, Bayesian analysis-via simulation studies. Real data on contact wrestling are used for illustration.

Active temperature compensation for an accelerometer based angle measuring device

An angle measuring device using a high performance and very compact accelerometer provides a new and exciting method for producing highly compact and accurate angle measuring devices. Accelerometers are micro-machined and are able to measure acceleration to a very high accuracy. By using gravity as a reference these compact devices can also be used for measuring angles of rotation. The inherent problem with these devices is that their response characteristic changes with temperature which is detrimental to measurement accuracy. This paper describes an effective method to overcome this problem using a temperature sensor and intelligent software to compensate for this drift characteristic. In order to demonstrate the effectiveness of this work, experiments have been developed and conducted with the results and analysis provided at the end
of this paper for discussion.

Effect of alloying additions on the SFE, Neél temperature and shape memory effect in Fe–Mn–Si-based alloys

A range of Fe–Mn–Si-based shape memory alloys has been investigated to examine the interplay of composition, stacking fault probability (SFP) and Neél temperature on the shape memory effect (SME). It has been found that the SFP (inversely proportional to stacking fault energy) showed little correlation to the SME for the range of alloy compositions examined. Further, the Neél temperature was not found to exhibit a significant effect on the SME. The addition of interstitial elements, however, was found to markedly decrease the SME.

Evaluation of the robustness of predictive yield models based on catchment characteristics using GIS for reservoir fisheries in Sri Lanka

Land-use patterns in the catchment areas of Sri Lankan reservoirs, which were quantified using Geographical Information Systems (GIS), were used to develop quantitative models for yield prediction. The validity of these models was evaluated through the application to five reservoirs that were not used in the development of the models, and by comparing with the actual fish yield data of these reservoirs collected by an independent body. The robustness of the predictive models developed was tested by principal component analysis (PCA) on limnological characteristics, land-use patterns of the catchments and fish yields. The predicted fish yields in five Sri Lankan reservoirs, using the empirical models based on the ratios of forest cover and/or shrub cover to reservoir capacity or reservoir area were in close agreement with the observed fish yields. The scores of PCA ordination of productivity-related limnological parameters and those of land-use patterns were linearly related to fish yields. The relationship between the PCA scores of limnological characteristics and land-use types had the appropriate algebraic form, which substantiates the influence of the limnological factors and land-use types on reservoir fish yields. It is suggested that the relatively high predictive power of the models developed on the basis of GIS methodologies can be used for more accurate assessment of reservoir fisheries. The study supports the importance and the need for an integrated management strategy for the whole watershed to enhance fish yields.

Rule-based dependency models for security protocol analysis

Security protocol analysis has been discussed for quite some time in the past few years. Although formal methods have been widely used to identify various vulnerabilities, mainly susceptibility to freshness attacks and impersonation, the arisen inconsistent data between principals and collusion attacks held by a group of dishonest principals have been largely ignored. Moreover, the previous methods focus on reasoning about certain security-related properties and detecting known attacks against secure message, whereas there have been insufficient efforts to handle the above hidden but powerful attacks. In this paper, we address these critical issues and prove the efficiency and intuitiveness of rule-based dependency models in defending a protocol against the attacks. This is able to provide a numerical estimation to measure he occurrence of these attacks. It will be useful in enhancing the current protocol analysis.

Laboratory and non-laboratory-based risk prevention models for secondary prevention of cardiovascular disease : the LIPID study

Aims : The aims of this study were to examine whether risk prediction models for recurrent cardiovascular disease (CVD) events have prognostic value, and to particularly examine the performance of those models based on non-laboratory data. We also aimed to construct a risk chart based on the risk factors that showed the strongest relationship with CVD.

Methods and results : Cox proportional hazards models were used to calculate a risk score for each recurrent event in a CVD patient who was enrolled in a very large randomized controlled clinical trial. Patients were then classified into groups according to quintiles of their risk score. These risk models were validated by calibration and discrimination analyses based on data from patients recruited in New Zealand for the same study. Non-laboratory-based risk factors, such as age, sex, body mass index, smoking status, angina grade, history of myocardial infarction, revascularization, stroke, diabetes or hypertension and treatment with pravastatin, were found to be significantly associated with the risk of developing a recurrent CVD event. Patients who were classified into the medium and high-risk groups had two-fold and four-fold the risk of developing a CVD event compared with those in the low-risk group, respectively. The risk prediction models also fitted New Zealand data well after recalibration.

Conclusion : A simpler non-laboratory-based risk prediction model performed equally as well as the more comprehensive laboratory-based risk prediction models. The risk chart based on the further simplified Score Model may provide a useful tool for clinical cardiologists to assess an individual patient's risk for recurrent CVD events.

Recognition of idle resources in time-driven activity-based costing and resource consumption accounting models

Despite its theoretical superiority over traditional volume-based costing models, the Activity-Based Costing (ABC) model has failed to replace traditional volume-based costing models in most organisations. In response to the problems of the model, Time-Driven Activity-Based Costing (TDABC) and Resource Consumption Accounting (RCA) models have been developed as costing models for next generation cost management systems. A key feature that distinguishes TDABC and RCA models from traditional volume-based costing models and the ABC model is the recognition of idle resources in resource pools.

This paper presents a discussion on implications of recognising idle resources in TDABC and RCA models on developments, maintenance and uses of cost management systems. A hypothetical case is presented to illustrate conversions of an ABC-based costing model to ones that are based on the TDABC and RCA models, and the resulting new allocation of resource costs.

Lithium-polymer battery based on an ionic liquid–polymer electrolyte composite for room temperature applications

A lithium-polymer battery based on an ionic liquid–polymer electrolyte (IL–PE) composite membrane operating at room temperature is described. Utilizing a polypyrrole coated LiV₃O₈ cathode material, the cell delivers >200 mAh g⁻¹ with respect to the mass of the cathode material. Discharge capacity is slightly higher than those observed for this cathode material in standard aprotic electrolytes; it is thought that this is the result of a lower solubility of the LiV₃O₈ material in the IL–PE composite membrane.

Room-temperature molten salts based on the quaternary ammonium ion

The properties of a family of novel quaternary ammonium salts based on the bis(trifluoromethylsulfonyl)imide and triflate anions are reported. Binary phase diagrams for some of their mixtures and their electrochemical windows of stability are also reported. The highest conductivity observed in the pure salt systems at 25 °C was 7 × 10^-4 S cm^-1. An electrochemical window of stability of up to 5 V was measured on graphite electrodes. The effect of salt structure and solvent on conductivity of the salts is also discussed.

Synthesis and properties of ambient temperature molten salts based on the quaternary ammonium ion

The synthesis of 16 tetraalkyl ammonium bis(trifluoromethane sulfonyl) imide salts, (C_nH_2n+1)₄ +N -N (SO₂CF₃)₂ (n = 1, 2, 3, 4),  (C₂H₅)₂(i-C₃H₇)₂ +N -N(SO₂CF₃)₂, (C₂H₅)(CH₃)(i-C₃H₇)₂+N-N(SO₂CF₃)₂, (n-C₇H₁₅)(C₂H₅)i-C₃H₇)₂+N-N(SO₂CF₃)₂ and (C_nH_2n+1)(C_mH_2m+1)₃+N-N(SO₂CF₃)₂ (n = 6,7,8; m = 1, 2, 4) are reported in this paper. Trends in properties of these salts are discussed. The symmetrical tetraalkyl ammonium salts with the bis(trifluoromethyl sulfonyl) imide anion exhibited a lower melting point than that of corresponding ammonium halides. The salts with low symmetry ammonium cations were found to be of generally lower melting point, and many were stable liquids at room temperature. Several of these did not crystallize during cooling below room temperature and exhibited glass transition temperatures in the region of −60 °C∼−80 °C. A comparison of properties between the ammonium imide salts and corresponding trifluoromethane sulfonates is also presented.

Suitability of chlorine bulk decay models for planning and management of water distribution systems

Effective disinfection planning and management in large, complex water distribution systems requires an accurate network water quality model. This model should be based on reaction kinetics, which describes disinfectant loss from bulk water over time, within experimental error. Models in the literature were reviewed for their ability to meet this requirement in real networks. Essential features were identified as accuracy, simplicity, computational efficiency, and ability to describe consistently the effects of initial chlorine dose, temperature variation, and successive rechlorinations. A reaction scheme of two organic constituents reacting with free chlorine was found to be necessary and sufficient to provide the required features. Recent release of the multispecies extension (MSX) to EPANET and MWH Soft's H2OMap Water MSX network software enables users to implement this and other multiple-reactant bulk decay models in real system simulations.