Using data mining for prediction of hospital length of stay: an application of the CRISP-DM Methodology

Hospitals are nowadays collecting vast amounts of data related with patient records. All this data hold valuable knowledge that can be used to improve hospital decision making. Data mining techniques aim precisely at the extraction of useful knowledge from raw data. This work describes an implementation of a medical data mining project approach based on the CRISP-DM methodology. Recent real-world data, from 2000 to 2013, were collected from a Portuguese hospital and related with inpatient hospitalization. The goal was to predict generic hospital Length Of Stay based on indicators that are commonly available at the hospitalization process (e.g., gender, age, episode type, medical specialty). At the data preparation stage, the data were cleaned and variables were selected and transformed, leading to 14 inputs. Next, at the modeling stage, a regression approach was adopted, where six learning methods were compared: Average Prediction, Multiple Regression, Decision Tree, Artificial Neural Network ensemble, Support Vector Machine and Random Forest. The best learning model was obtained by the Random Forest method, which presents a high quality coefficient of determination value (0.81). This model was then opened by using a sensitivity analysis procedure that revealed three influential input attributes: the hospital episode type, the physical service where the patient is hospitalized and the associated medical specialty. Such extracted knowledge confirmed that the obtained predictive model is credible and with potential value for supporting decisions of hospital managers.

An integrated system for human action recognition from video using hidden Markov model

In this paper, we present an integrated system for real-time automatic detection of human actions from video. The proposed approach uses the boundary of humans as the main feature for recognizing actions. Background subtraction is performed using Gaussian mixture model. Then, features are extracted from silhouettes and Vector Quantization is used to map features into symbols (bag of words approach). Finally, actions are detected using the Hidden Markov Model. The proposed system was validated using a newly collected real- world dataset. The obtained results show that the system is capable of achieving robust human detection, in both indoor and outdoor environments. Moreover, promising classification results were achieved when detecting two basic human actions: walking and sitting.

Dispersion and re-agglomeration of graphite nanoplates in polypropylene melts under controlled flow conditions

The kinetics of GnP dispersion in polypropylene melt was studied using a prototype small scale modular extensional mixer. Its modular nature enabled the sequential application of a mixing step, melt relaxation, and a second mixing step. The latter could reproduce the flow conditions on the first mixing step, or generate milder flow conditions. The effect of these sequences of flow constraints upon GnP dispersion along the mixer length was studied for composites with 2 and 10 wt.% GnP. The samples collected along the first mixing zone showed a gradual decrease of number and size of GnP agglomerates, at a rate that was independent of the flow conditions imposed to the melt, but dependent on composition. The relaxation zone induced GnP re-agglomeration, and the application of a second mixing step caused variable dispersion results that were largely dependent on the hydrodynamic stresses generated.

Comparison of different numerical methods for the solution of the time-fractional reaction-diffusion equation with variable diffusion coefficient

In this work we perform a comparison of two different numerical schemes for the solution of the time-fractional diffusion equation with variable diffusion coefficient and a nonlinear source term. The two methods are the implicit numerical scheme presented in [M.L. Morgado, M. Rebelo, Numerical approximation of distributed order reaction- diffusion equations, Journal of Computational and Applied Mathematics 275 (2015) 216-227] that is adapted to our type of equation, and a colocation method where Chebyshev polynomials are used to reduce the fractional differential equation to a system of ordinary differential equations

Connecting free volume with shape memory properties in noncytotoxic gamma-irradiated polycyclooctene

The free volume holes of a shape memory polymer have been analysed considering that the empty space between molecules is necessary for the molecular motion, and the shape memory response is based on polymer segments acting as molecular switches through variable flexibility with temperature or other stimuli. Therefore, thermomechanical analysis (TMA) and positron annihilation lifetime spectroscopy (PALS) have been applied to analyse shape recovery and free volume hole sizes in gamma irradiated polycyclooctene (PCO) samples, as a non-cytotoxic alternative to more conventional PCO crosslinked via peroxide for future applications in medicine. Thus, a first approach relating structure, free volume holes and shape memory properties in gamma irradiated PCO is presented. The results suggest that free volume holes caused by gamma irradiation in PCO samples facilitate the recovery process by improving movement of polymer chains and open t possibilities for the design and control of the macroscopic response.

Risk of multiple myeloma is associated with polymorphisms within telomerase genes and telomere length

Compelling biological and epidemiological evidences point to a key role of genetic variants of the TERT and TERC genes in cancer development. We analyzed the genetic variability of these two gene regions using samples of 2,267 multiple myeloma (MM) cases and 2,796 healthy controls. We found that a TERT variant, rs2242652, is associated with reduced MM susceptibility (OR?=?0.81; 95% CI: 0.72-0.92; p?=?0.001). In addition we measured the leukocyte telomere length (LTL) in a subgroup of 140 cases who were chemotherapy-free at the time of blood donation and 468 controls, and found that MM patients had longer telomeres compared to controls (OR?=?1.19; 95% CI: 0.63-2.24; ptrend ?=?0.01 comparing the quartile with the longest LTL versus the shortest LTL). Our data suggest the hypothesis of decreased disease risk by genetic variants that reduce the efficiency of the telomerase complex. This reduced efficiency leads to shorter telomere ends, which in turn may also be a marker of decreased MM risk.

On the length of bank-firm relationships: an empirical application to a major French bank

NIPE - WP 01/ 2016

Effect of the alkyl chain length of the ionic liquid anion on polymer electrolytes properties

New polymer electrolytes (PEs) based on chitosan and three ionic liquid (IL) families ([C2mim][CnSO3], [C2mim][CnSO4] and [C2mim][diCnPO4]) were synthesized by the solvent casting method. The effect of the length of the alkyl chain of the IL anion on the thermal, morphological and electrochemical properties of the PEs was studied. The solid polymer electrolytes (SPE) membranes were analyzed by differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive X-ray (EDX), polarized optical microscopy (POM), atomic force microscopy (AFM), complex impedance spectroscopy (ionic conductivity) and cyclic voltammetry (CV). The obtained results evidenced an influence of the alkyl chain length of the IL anion on the temperature of degradation, birefringence, surface roughness and ionic conductivity of the membranes. The DSC, XRD and CV results showed independency from the length of the IL-anion-alkyl chain. The PEs displayed an predominantly amorphous morphology, a minimum temperature of degradation of 135 °C, a room temperature (T = 25 °C) ionic conductivity of 7.78 × 10−4 S cm−1 and a wide electrochemical window of ∼ 4.0 V.

Nanocomposites: the contribution of multiscale modelling to improve the dispersion levels

Tese de Doutoramento em Ciência e Engenharia de Polímeros e Compósitos

Predicting SO2 pollution incidents by means of additive models with optimum variable selection

The aim of this paper is to predict time series of SO2 concentrations emitted by coal-fired power stations in order to estimate in advance emission episodes and analyze the influence of some meteorological variables in the prediction. An emission episode is said to occur when the series of bi-hourly means of SO2 is greater than a specific level. For coal-fired power stations it is essential to predict emission epi- sodes sufficiently in advance so appropriate preventive measures can be taken. We proposed a meth- odology to predict SO2 emission episodes based on using an additive model and an algorithm for variable selection. The methodology was applied to the estimation of SO2 emissions registered in sampling lo- cations near a coal-fired power station located in Northern Spain. The results obtained indicate a good performance of the model considering only two terms of the time series and that the inclusion of the meteorological variables in the model is not significant.