Flexible Distribution-Based Regression Models for Count Data: Application to Medical Diagnosis

Abstract

Data mining techniques have been successfully utilized in different applications of significant fields, including medical research. With the wealth of data available within the health-care systems, there is a lack of practical analysis tools to discover hidden relationships and trends in data. The complexity of medical data that is unfavorable for most models is a considerable challenge in prediction. The ability of a model to perform accurately and efficiently in disease diagnosis is extremely significant. Thus, the model must be selected to fit the data better, such that the learning from previous data is most efficient, and the diagnosis of the disease is highly accurate. This work is motivated by the limited number of regression analysis tools for multivariate counts in the literature. We propose two regression models for count data based on flexible distributions, namely, the multinomial Beta-Liouville and multinomial scaled Dirichlet, and evaluated the proposed models in the problem of disease diagnosis. The performance is evaluated based on the accuracy of the prediction which depends on the nature and complexity of the dataset. Our results show the efficiency of the two proposed regression models where the prediction performance of both models is competitive to other previously used regression models for count data and to the best results in the literature.     

Ultimate bearing capacity of strip footing resting on clay soil mixed with tire-derived aggregates

This study investigated the use of recycled tire-derived aggregate (TDA) mixed with kaolin as a method of increasing the ultimate bearing capacity ( UBC) of a strip footing. Thirteen 1g physical modeling tests were prepared in a rigid box of 0.6 m × 0.9 m in plan and 0.6 m in height. During sample preparation, 0%, 20%, 40%, or 60% (by weight) of powdery, shredded, small-sized granular (G 1–4 mm) or large-sized granular (G 5–8 mm) TDA was mixed with the kaolin. A strip footing was then placed on the stabilized kaolin and was caused to fail under stress-controlled conditions to determine the UBC. A rigorous 3D finite element analysis was developed in Optum G-3 to determine the UBC values based on the experimental test results. The experimental results showed that, except for the 20% powdery TDA, the TDA showed an increase in the UBC of the strip footing. When kaolin mixed with 20% G (5–8 mm), the UBC showed a threefold increase over that for the unreinforced case. The test with 20% G (1–4 mm) recorded the highest subgrade modulus. It was observed that the UBC calculated using finite element modeling overestimated the experimental UBC by an average of 9%.     

Al2O3/Yttria‐Stabilized Zirconia Hollow‐Fiber Membrane Incorporated with Iron Oxide for Pb(II) Removal

Removal by absorptive ceramic membranes can simultaneously absorb and separate metal ions from water. Alumina/yttria‐stabilized zirconia (Al₂O₃/YSZ) hollow‐fiber membranes, fabricated using phase inversion and sintering process, were deposited with iron oxide by an in‐situ hydrothermal process. The results showed that α‐Fe₂O₃ was produced and incorporated across the membranes. A reduction in flux was recorded with the deposition of α‐Fe₂O_3. However, it improved the adsorption capacity for heavy metal adsorption. The adsorption‐separation test demonstrated that the optimized membrane is able to completely remove Pb(II) ions after two hours.     

Mobility management support and performance analysis for wireless MPLS networks

Multi‐Protocol Label Switching (MPLS) is deployed in the Internet backbone to support service differentiation and traffic engineering. In recent years, there has been interest in extending the MPLS capability to wireless access networks for mobility management support. In this paper, we present analysis of Micro Mobile MPLS, a new micro‐mobility management scheme which integrates the Mobile IP and MPLS protocols by using two‐level hierarchy architecture. Our proposal supports two protocol variants. First, the fast handoff process, which anticipates the LSP procedure set‐up with neighboring locations where a mobile node (MN) may move to, is provided to reduce service disruption. Second, a new mechanism based on the forwarding chain concept is proposed to track efficiently the host mobility within a domain. This concept can significantly reduce registration update costs and provide low handoff latency. Analytical models are developed and simulations are conducted to justify the benefits of our proposed mechanisms. Copyright © 2006 John Wiley & Sons, Ltd.     

SO2 SCRUBBING BY ULTRAFINE CA(OH)2 AEROSOL

The objective of the present study was to generate submicrometer calcium hydroxide aerosols and to investigate the effectiveness of such aerosols in sulfur capture. The effectiveness of SO₂ removal by Ca(OH)₂ aerosol has been investigated in an isothermal reactor. Ca(OH) ₂ aerosol was generated by a novel fluidizer system in which submicrometer-sized powders were entrained in gases. SO₂ was added to this aerosol to a concentration of 2000 ppm. The aerosol-SO₂ mixture was heated to 550°C-750°C in an isothermal tube reactor. The SO₂ removal efficiency, which varied from 20% to 70%, was determined to be a function of the aerosol concentration, reactor temperature and residence time. The fraction of aerosol reacted was not affected strongly by the aerosol concentration. The reaction kinetics were determined from the experimental data using a simple analytical model in which the rate is first order in both SO₂ and calcium hydroxide aerosol concentrations.     

Graft Copolymerization of Acrylamide onto Oil Palm Empty Fruit Bunch (OPEFB) Fiber

Grafting of acrylamide (AAm) onto oil palm empty fruits bunch fiber using hydrogen peroxide as initiator and methyl acrylate as comonomer was investigated. The amount of comonomer needed to make grafting of acrylamide possible was determined. The percentage of poly(acrylamide) and the comonomer in the final graft copolymer was estimated by elemental analysis. Results obtained indicated that methyl acrylate facilitated the incorporation of acrylamide monomer onto OPEFB. The reactivity ratios for both monomers were determined by using Fineman–Ross plot. The effects of reaction temperature and period as well as amount of the initiator, solvent, monomer and comonomer on the percentage of grafting at fixed amount of comonomer (11 mmol) were studied. Maximum percentage of grafting was achieved when the amount of initiator and solvent 3.98×10⁻³ mol and 50 mL respectively. The optimum reaction temperature was 50 ^○C and the reaction period was 90 min. Highest percentage of grafting was 232% when 25.6 mmol of acrylamide was used under these optimum conditions. The presence of functional group in the grafted polymer is characterized by infrared spectroscopy and the surface morphology is observed by scanning electron microscopy. Thermoanalytic investigation on OPEFB and OPEFB-g-PAAM were carried out to evaluate the thermal stability and respective activation energy of the materials.     

Variational learning for Dirichlet process mixtures of Dirichlet distributions and applications

In this paper, we propose a Bayesian nonparametric approach for modeling and selection based on a mixture of Dirichlet processes with Dirichlet distributions, which can also be seen as an infinite Dirichlet mixture model. The proposed model uses a stick-breaking representation and is learned by a variational inference method. Due to the nature of Bayesian nonparametric approach, the problems of overfitting and underfitting are prevented. Moreover, the obstacle of estimating the correct number of clusters is sidestepped by assuming an infinite number of clusters. Compared to other approximation techniques, such as Markov chain Monte Carlo (MCMC), which require high computational cost and whose convergence is difficult to diagnose, the whole inference process in the proposed variational learning framework is analytically tractable with closed-form solutions. Additionally, the proposed infinite Dirichlet mixture model with variational learning requires only a modest amount of computational power which makes it suitable to large applications. The effectiveness of our model is experimentally investigated through both synthetic data sets and challenging real-life multimedia applications namely image spam filtering and human action videos categorization.     

Rheological,thermal, and morphological properties of blends based on poly(propylene), ethylene propylene rubber,and ethylene-1-octene copolymer that could result from end of life vehicles: Effect of maleic anhydride grafted poly(propylene)

The objective of this work is to study the properties of blends that could result from the recycling of end-of-life vehicles (ELV). While ethylene propylene rubber (EPR) and ethylene propylene diene monomer (EPDM) have been used extensively as elastomeric additives in poly(propylene) (PP), they can be substituted by ethylene-1-octene copolymer (EOC). As a consequence, the matter resulting from the sorting of ELV might be more complex and made of PP, EPR, and EOC. The effect of incorporating EOC [that is a polyethylene elastomer (PEE)] and maleic anhydride grafted polypropylene (PP-g-MAH) on the rheological, thermal, and morphological properties of PP/EPR blends has been investigated. Blends of various compositions (with and without compatibilizer) were prepared using a corotating twin-screw extruder. The results were compared to the ones presented by a commercial (PP/EPR) blend. The EPR phase is dispersed in the form of spherical particles in (PP/EPR). The EOC phase is dispersed in the form of aggregated particles. Dynamic viscoelastic and differential scanning calorimetry properties of (PP/EPR)/EOC blends shows the incompatibility of the components even in presence of PP-g-MAH copolymer. POLYM. ENG. SCI., 47:1009–1015, 2007. © 2007 Society of Plastics Engineers     

A statistical framework for few-shot action recognition

Along with the exponential growth of online video creation platforms such as Tik Tok and Instagram, state of the art research involving quick and effective action/gesture recognition remains crucial. This work addresses the challenge of classifying short video clips, using a domain-specific feature design approach, capable of performing significantly well using as little as one training example per action. The method is based on Gunner Farneback’s dense optical flow (GF-OF) estimation strategy, Gaussian mixture models, and information divergence. We first aim to obtain accurate representations of the human movements/actions by clustering the results given by GF-OF using K-means method of vector quantization. We then proceed by representing the result of one instance of each action by a Gaussian mixture model. Furthermore, using Kullback-Leibler divergence (KL-divergence), we attempt to find similarities between the trained actions and the ones in the test videos. Classification is done by matching each test video to the trained action with the highest similarity (a.k.a lowest KL-divergence). We have performed experiments on the KTH and Weizmann Human Action datasets using One-Shot and K-Shot learning approaches, and the results reveal the discriminative nature of our proposed methodology in comparison with state-of-the-art techniques.

     

Color object segmentation and tracking using flexible statistical model and level-set

Multimedia Tools and Applications - This study presents an unsupervised novel algorithm for color image segmentation, object detection and tracking based on unsupervised learning step followed with...