| No. of records per page: 10 | 20 | 30 | 50 | 100 | Show all |
Record ID: 50 [ Page 12 of 16, No. 1 ]
Modelling Zero-Inflated Clustered Count Data: A Semiparametric Approach
Authors: Kevin Carl P. Santos
Abstract:
This paper proposes to use an additive semiparametric Poisson regression in modelling zero-inflated clustered data. Two estimation methods are exploited in this paper based on de Vera (2010). The first simultaneously estimates both the parametric and nonparametric parts of the model. The second utilizes the backfitting algorithm by smoothing the nonparametric function of the covariates and then estimating the parametric parts of the postulated model. The predictive accuracy, measured in terms of root mean square error (RMSE), of the proposed methods is compared to that of ordinary Zero-Inflated Poisson (ZIP) regression model. It is found out through simulation study that the average RMSE of the ordinary ZIP regression model is at most 81% and 27% higher for equal and unequal cluster sizes, respectively, than that of proposed model whose parametric and nonparametric parts are simultaneously estimated.
Keywords: Zero-Inflated Poisson models, clustered data, Generalized Additive Models, backfitting algorithm
Download this article:
Year: 2014 Vol.: 63 No.: 1
Record ID: 49 [ Page 12 of 16, No. 2 ]
Autologistic Spatial-Temporal Modeling
Authors: Ma. Andriena Ida B. Del Ayre-Ofina
Abstract:
We postulate a combination of spatial-temporal and autologistic model in characterizing binary data collected over time and space. Using a second-order neighborhood system in defining the spatial component of the model, backfitting algorithm is used in estimating the model. As the incidence of success and failure responses becomes balanced, sensitivity and specificity increases. The predictive ability of the model is fairly robust to the spatial parameter but is significantly influenced by the temporal parameter. The bias of the estimate for the spatial parameter declines as it becomes dominant into the model. Furthermore, as the autocorrelation becomes stronger, its estimate becomes less biased. The backfitting algorithm is also observed to converge fast in the estimation of the spatial-temporal autologistic model.
Keywords: binary response, autologistic model, spatial-temporal model, backfitting
Download this article:
Year: 2014 Vol.: 63 No.: 1
Record ID: 48 [ Page 12 of 16, No. 3 ]
Visual Exploration of Climate Variability
Authors: Wendell Q. Campano; Rona Mae U. Tadlas
Abstract:
In this paper, a data visualization framework for investigating and exploring climate time series data is introduced. This method utilizes the results obtained from performing series of cluster analysis based on a particular multivariate data set for each defined subset in the time series. The said approach is implemented to the climate data in the Philippines. The data image results obtained from the procedure revealed the expected overall climate pattern in the Philippines as well as some localized segments of climate changes in the time series which deviate from the overall pattern. A wavelet analysis which is a well established method in analyzing climate data is also done to validate the results shown by the proposed visualization method.
Keywords: information visualization; data image; cluster analysis; wavelet; climate change; climate variability; time series; multivariate data
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 47 [ Page 12 of 16, No. 4 ]
Measuring Income Mobility using Pseudo-Panel Data
Authors: Arturo M. Martinez Jr; Mark Western; Michele Haynes; Wojtek Tomaszewski
Abstract:
To reconcile the need of providing a more dynamic perspective of the evolution of income distribution with the lack of panel data, several techniques have been offered to construct pseudo-panel data from repeated cross-sectional surveys. Using actual panel data from the Philippines, this study evaluates the performance of four pseudo-panel techniques in measuring a wide array of income mobility indicators. Preliminary results suggest that methods with more flexible income model specifications perform better than those with highly parameterized models. More importantly, these flexible pseudo-panel procedures produced estimates of poverty dynamics and movement-based indices which are quite close to the estimates computed from the actual panel data. Nevertheless, further improvements are warranted to be able to develop a more satisfactory estimation procedure for indices measuring temporal dependence and the inequality-reducing effect of income mobility.
Keywords: panel survey; cross-sectional survey; temporal; dependence; income distribution
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 46 [ Page 12 of 16, No. 5 ]
Effects of Education on Climate Risk Vulnerability in the Philippines: Evidence from Regional Panel Data
Authors: Michael Daniel C. Lucagbo; Kristina Norma B. Cobrador; Nikki Ann M. de Mesa; Remy Faye M. Ferrera; Jennifer E. Marasigan
Abstract:
The effects of climate change are being felt disproportionately in the world’s poorest countries and among those groups of people least able to cope. The Philippines, being a storm-lashed nation, is one country having high climate change vulnerability and low climate change resilience. A number of researches have suggested investments on adaptation which place strong emphasis on reducing vulnerability to climate change. Focusing on climate change vulnerability in the Philippines, this study examines the effect of one particular type of government intervention: increasing the level of education. In this study, the effect of education on vulnerability to climate change is examined in a regional panel data analysis using official Philippine statistics from the Natural Disaster Risk Reduction and Management Council (NDRRMC), Labor Force Survey (LFS), National Statistical Coordination Board (NSCB). Using the fixed-effects Poisson (FEP) regression model, the study establishes that at the community level, the number of employed college graduates is a significant factor that reduces climate risk vulnerability (measured by a number of deaths from natural disasters), controlling for other factors such as number of disasters, gross regional domestic product (GRDP), and population size.
Keywords: Vulnerability, Resilience, Panel Data, Fixed-effects Poisson model
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 45 [ Page 12 of 16, No. 6 ]
Regression Analyses of the Philippine Birth Weight Distribution
Authors: Elline Jade Beltran; Robert Neil F. Leong; Frumencio F. Co
Abstract:
Low birth weight has both short-term and long-term effects. It can lead to complications among infants causing neonatal deaths. Several literatures also suggested relationships between low birth weight and delayed mental and physical development. These negative effects are further magnified in developing countries, one of which is the Philippines. In this paper, birth weight is analysed through logistic, ordinary least squares, and quantile regression techniques using a sample from the 2008 Philippine Birth Recode. Quantile regression results offer a more dynamic picture of how these correlates affect the conditional distribution of birth weight. The obtained estimates of the marginal effects of several demographical and maternal health correlates of birth weight suggest that socially and economically impoverished mothers are more likely to have low birth weight babies. These results would recommend a focus on improving maternal health care through proper education.
Keywords: birth weight; quantile regression; logistic regression; ordinary least squares
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 44 [ Page 12 of 16, No. 7 ]
Profitability and Growth Topology Analysis of Unilevel-type of Network Marketing Structures
Authors: John Carlo P. Daquis; Angelique O. Castaneda; Nelson D. Sy; Joseph V. Abgona
Abstract:
This study analyzes a type of multi-level marketing (MLM) structure through a simulation of MLM systems. In unilevel MLM, distributors earn from both sales from direct selling and commissions from recruitment of downlines. Several distributional assumptions were made in constructing the system, such as the use of the uniform, Bernoulli, and Poisson distributions. Member income is measured based on commission from recruit pay-ins in their downlines and income from direct selling. Based on the simulated unilevel MLM structures, the fundamental behavior of a unilevel MLM is captured and analyzed in terms of its network growth topology and profitability.
Keywords: multi-level marketing; network simulation; unilevel structure; complex systems; probability distributions
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 43 [ Page 12 of 16, No. 8 ]
Classification of Congenital Hypothyroidism using Artificial Neural Networks
Authors: Iris Ivy Gauran; Ma. Sofia Criselda A. Poblador
Abstract:
The Newborn Screening Reference Center (NSRC) of the National Health Institute in University of the Philippines Manila collects measurements from five attributes to determine whether Congenital Hypothyroidism (CH) is present in a neonate. Detecting the CH cases is a major concern of medical practitioners because it provides richer information than the healthy ones. However, because of the rarity of this metabolic condition, existing classification algorithms oftentimes misclassify a newborn as “normal” even if it is not. This paper investigates the efficiency of Self-Organizing Kohonen Maps (SOM), a type of artificial neural network. Though it is a visualization and clustering tool, the researchers want to probe on its ability to detect outliers and properly classify a newborn as normal or not by coming up with a statistically computed threshold value. Instead of working directly with the original attributes of the data, a reduced set of SOM prototypes is utilized to represent the data in a space of smaller dimension, seeking to preserve the probability distribution and topology of the input space. Results showed a misclassification rate of 13.5%. Though it is found to be slightly less superior to the existing classification rules, the proposed methodology was able to address the problem of finding a statistical threshold value. Also, the methodology verifies that age has a major effect on misclassifying “Normal” as “Abnormal” since postponement of newborn screening to a later age causes the quantization error to boost drastically, hence, easily exceeding the value of the first decision threshold.
Keywords: self-organizing kohonen maps (SOM), classification algorithm, outlier detection, newborn screening for congenital hypothyroidism
Download this article:
Year: 2013 Vol.: 62 No.: 2
Record ID: 42 [ Page 12 of 16, No. 9 ]
Career opportunities in the pharmaceutical industry
Authors: Jennifer Ly
Abstract:
Keywords:
Download this article:
Year: 2013 Vol.: 62 No.: 1
Record ID: 41 [ Page 12 of 16, No. 10 ]
An elementary proof of independence of least squares estimation of regression coefficients and of variance in linear regression
Authors: Alexaander R. De Leon; Joyce Raymund B. Punzalan
Abstract:
Keywords:
Download this article:
Year: 2013 Vol.: 62 No.: 1