Anwar Fitrianto

Panel Data Analysis for Sabah Construction Industries: Choosing the Best Model☆

Analysis of panel data by using statistical models is rapidly growing. It is sometime tough for the novice users of panel data to make an informed choice of what estimators best suit their research questions. This paper is meant to find best model among few types of models such as panel data models and ordinary least squares (OLS) regression for Sabah construction industries. The best model will be chosen based on lowest Root Mean Square Errors (RMSE). The purpose of comparing between models is to find the most efficient model which will be useful for prediction. After analyzing the data using SAS software, it was found that two-way fixed effect panel data model provide the lowest RMSE for the Sabah construction industries.

Modified boxplot for extreme data

A boxplot is an exploratory data analysis (EDA) tool for a compact distributional summary of a data set. It is designed to captures all typical observations and displays the location, spread, skewness and the tail of the data. The precision of some of this functionality is considered to be more reliable for symmetric data type and thus less appropriate for skewed data such as the extreme data. Many observations from extreme data were erroneously marked as outliers by the Tukeys standard boxplot. We proposed a modified boxplot fence adjustment using the Bowley coefficient, a robust skewness measure. The adjustment will enable us to detect inconsistent observations without any parametric assumption about the distribution of the data. The new boxplot is capable of displaying some additional features such as the location parameter region of the Gumbel fitted extreme data. A simulated and real life data were used to show the advantages of this development over those found in the literature.

Development of Alginate - Gum Arabic Beads for Targeted Delivery of Protein

Controlled release beads were prepared by using the combination of alginate and gum Arabic through ionotropic gelation method. Bovine serum albumin was used as model protein for in vitro assessments. The effect of amount of sodium alginate and gum Arabic as the factor affecting protein encapsulation efficiency and protein release were optimized and analyzed by using RSM-FCCD. It was observed that protein encapsulation efficiency was increased and protein release was decreased with the increase of both of the amount of sodium alginate and gum Arabic, used as polymer blend. The optimized beads showed high encapsulation efficiency (87.5 ± 3.65%) with suitable protein release (100% protein release after almost 4 hrs). The swelling of beads were highly influenced by pH of dissolution medium. These beads were also characterized by FT-IR spectroscopy, SEM and TA for protein-excipients interaction, beads surface morphology and beads strength, respectively. These calcium alginate/gum Arabic beads have good potential to be used as delivery vehicle for protein drugs.

A Comparative Study of Linear and Nonlinear Regression Models for Outlier Detection

Artificial Neural Networks provide models for a large class of natural and artificial phenomena that are difficult to handle using classical parametric techniques. They offer a potential solution to fit all the data, including any outliers, instead of removing them. This paper compares the predictive performance of linear and nonlinear models in outlier detection. The best-subsets regression algorithm for the selection of minimum variables in a linear regression model is used by removing predictors that are irrelevant to the task to be learned. Then, the ANN is trained by the Multi-Layer Perceptron to improve the classification and prediction of the linear model based on standard nonlinear functions which are inherent in ANNs. Comparison of linear and nonlinear models was carried out by analyzing the Receiver Operating Characteristic curves in terms of accuracy and misclassification rates for linear and nonlinear models. The results for linear and nonlinear models achieved 68% and 93%, respectively, with better fit for the nonlinear model.

Modeling Asia's Child Mortality Rate: A Thinking of Human Development in Asia

Abstract Multiple linear regression model was employed to model child under age of five mortality rate and related factors in Asia of year 2010. Data analysis was carried out to find factors which influence the child mortality in Asia. Correlation analysis was done to check on the relationship among all the variables, as well as to identify the problem of multicollinearity in the data. Having fitted multiple linear regression, it was found that mortality rate of children under age of five in Asia countries are significantly influenced by percentage of case detection for all forms of tuberculosis, number of reported deaths on measles, number of population using an improved drinking water source, and number of birth trauma reported. Among those variable, it was identified that number of population using an improved drinking water source is the most important factor.

Identification of Multiple Outliers in a Generalized Linear Model with Continuous Variables

In the statistical analysis of data, a model might be awfully fitted with the presence of outliers. Besides, it has been well established to use residuals for identification of outliers. The asymptotic properties of residuals can be utilized to contribute diagnostic tools. However, it is now evident that most of the existing diagnostic methods have failed in identifying multiple outliers. Therefore, this paper proposed a diagnostic method for the identification of multiple outliers in GLM, where traditionally used outlier detection methods are effortless as they undergo masking or swamping dilemma. Hence, an investigation was carried out to determine the capability of the proposed GSCPR method. The findings obtained from the numerical examples indicated that the performance of the proposed method was satisfactory for the identification of multiple outliers. Meanwhile, in the simulation study, two scenarios were considered to assess the validity of the proposed method. The proposed method consistently displayed higher percentage of correct detection, as well as lower rates of swamping and masking, regardless of the sample size and the contamination levels.

IMPROVED NORMALIZATION AND STANDARDIZATION TECHNIQUES FOR HIGHER PURITY IN K-MEANS CLUSTERING

Clustering is basically one of the major sources of primary data mining tools, which make researchers understand the natural grouping of attributes in datasets. Clustering is an unsupervised classification method with aim of partitioning, where objects in the same cluster are similar, and objects belong to different clusters vary significantly, with respect to their attributes. The K-means algorithm is a famous and fast technique in non-hierarchical cluster algorithms. Based on its simplicity, the K-means algorithm has been used in many fields. This paper proposes improved normalization and standardization techniques for higher purity in K-means clustering experimented with benchmark datasets from UCI machine learning repository and it was found that all the proposed techniques’ performance was much higher compared to the conventional K-means and the three classic transformations, and it is evidently shown by purity and Rand index accuracy results.

Effects of a single outlier on the coefficient of determination: An empirical study

This article investigates the effects of outliers on the coefficient of determination, R2 which is computed by Ordinary Least Squares (OLS) estimator. It is now evident that the OLS is greatly affected by outliers and hence the R2 is also affected. This problem can be solved by using the robust estimators such as Least Trimmed Squares (LTS) estimator. In this article, we compare the value of R2 which is computed by OLS and LTS estimators. We modify a regression data set to effectively generate outliers in both X and Y directions. Then the coefficient of determination (OLS and LTS) is investigated from the modified data sets (data with outliers). The numerical results show the merit of using the LTS based R2 estimator compared to the OLS estimator.

On the Performance of Several Approaches to Obtain Standardized Logistic Regression Coefficients

In general regression analysis, standardized beta weights are often used to compare strength of prediction across variables. In order to consider obtaining an estimation of standardized logistic regression coefficients, the model must be rescaled to include such coefficients. One of the reasons to have the coefficient standardized is we will have more informative coefficients compared to unstandardized coefficients, especially for variables which have no natural metric. Several approaches of obtaining standardized coefficients in logistic regression are available in literatures. This article studies the performance of the existing approaches to obtain standardized logistic regression coefficient based on real data.

Estimating bias and variances in bootstrap logistic regression for Umaru and impact data

We employed random-x bootstrap in binary logistic regression model. We investigate the effect of sample size and number of bootstrap replication on the bias and variance. The performance of estimated coefficient is measured based on the bias, variance, and confidence interval of the bootstrap estimates. In addition, we also focus on the length of confidence interval of the bootstrap estimates. We found that bias and variance decrease for larger sample size. We noticed that length of confidence intervals decrease as the sample size and number of bootstrap replication are getting large. The results show that the estimated coefficient is more precise as the sample size increases.