|Commenced in January 2007||Frequency: Monthly||Edition: International||Paper Count: 6|
A major challenge in medical studies, especially those that are longitudinal, is the problem of missing measurements which hinders the effective application of many machine learning algorithms. Furthermore, recent Alzheimer's Disease studies have focused on the delineation of Early Mild Cognitive Impairment (EMCI) and Late Mild Cognitive Impairment (LMCI) from cognitively normal controls (CN) which is essential for developing effective and early treatment methods. To address the aforementioned challenges, this paper explores the potential of using the eXtreme Gradient Boosting (XGBoost) algorithm in handling missing values in multiclass classification. We seek a generalized classification scheme where all prodromal stages of the disease are considered simultaneously in the classification and decision-making processes. Given the large number of subjects (1631) included in this study and in the presence of almost 28% missing values, we investigated the performance of XGBoost on the classification of the four classes of AD, NC, EMCI, and LMCI. Using 10-fold cross validation technique, XGBoost is shown to outperform other state-of-the-art classification algorithms by 3% in terms of accuracy and F-score. Our model achieved an accuracy of 80.52%, a precision of 80.62% and recall of 80.51%, supporting the more natural and promising multiclass classification.
STRIM (Statistical Test Rule Induction Method) has been proposed as a method to effectively induct if-then rules from the decision table which is considered as a sample set obtained from the population of interest. Its usefulness has been confirmed by simulation experiments specifying rules in advance, and by comparison with conventional methods. However, scope for future development remains before STRIM can be applied to the analysis of real-world data sets. The first requirement is to determine the size of the dataset needed for inducting true rules, since finding statistically significant rules is the core of the method. The second is to examine the capacity of rule induction from datasets with contaminated attribute values created by missing data and noise, since real-world datasets usually contain such contaminated data. This paper examines the first problem theoretically, in connection with the rule length. The second problem is then examined in a simulation experiment, utilizing the critical size of dataset derived from the first step. The experimental results show that STRIM is highly robust in the analysis of datasets with contaminated attribute values, and hence is applicable to real-world data
Analyzing DNA microarray data sets is a great challenge, which faces the bioinformaticians due to the complication of using statistical and machine learning techniques. The challenge will be doubled if the microarray data sets contain missing data, which happens regularly because these techniques cannot deal with missing data. One of the most important data analysis process on the microarray data set is feature selection. This process finds the most important genes that affect certain disease. In this paper, we introduce a technique for imputing the missing data in microarray data sets while performing feature selection.
Ratio and regression type estimators have been used by previous authors to estimate a population mean for the principal variable from samples in which both auxiliary x and principal y variable data are available. However, missing data are a common problem in statistical analyses with real data. Ratio and regression type estimators have also been used for imputing values of missing y data. In this paper, six new ratio and regression type estimators are proposed for imputing values for any missing y data and estimating a population mean for y from samples with missing x and/or y data. A simulation study has been conducted to compare the six ratio and regression type estimators with a previous estimator of Rueda. Two population sizes N = 1,000 and 5,000 have been considered with sample sizes of 10% and 30% and with correlation coefficients between population variables X and Y of 0.5 and 0.8. In the simulations, 10 and 40 percent of sample y values and 10 and 40 percent of sample x values were randomly designated as missing. The new ratio and regression type estimators give similar mean absolute percentage errors that are smaller than the Rueda estimator for all cases. The new estimators give a large reduction in errors for the case of 40% missing y values and sampling fraction of 30%.