The Treatment of Missing Values in Discriminant Analysis—I. The Sampling Experiment

Chan, Linda S.; Dunn, Olive Jean

doi:10.1080/01621459.1972.10482414

Cited by 34 publications

(34 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pairwise deletion is typically more accurate (Gleason & Staelin, 1975;Kim & Curry, 1977) and regression imputation is most accurate (Beale & Little, 1975;Buck, 1960;Chan & Dunn, 1972;Gleason & Staelin, 1975;Raymond & Roberts, 1987). Mean substitution is problematic since inserting means in place of missing data tends to reduce variance in a given variable and hence reduce covariance with another variable.…”

Section: Missing Data Techniquesmentioning

confidence: 97%

The impact of four missing data techniques on validity estimates in human resource management

1996

View full text Add to dashboard Cite

ABSTRACT. While missing data are a common problem in field settings, there is relatively little information in Human Resource Management to guide researchers when they conduct analyses with incomplete data. This article discusses four techniques to deal with missing data. The implications of using listwise deletion, pairwise deletion, mean substitution, and regression estimation are demonstrated in an applied selection situation. The importance of the manner in which data were missing is analyzed and discussed.

show abstract

Section: Missing Data Techniquesmentioning

confidence: 97%

The impact of four missing data techniques on validity estimates in human resource management

1996

View full text Add to dashboard Cite

show abstract

“…Although these sophisticated methods exist for handling missing data, the effects of using some of the imputation methods on the performance of statistical discrimination algorithms remain unknown. Previously, Chan and Dunn [2] and Chan et al [3] have compared several ad hoc techniques to replace missing data, while Twedt and Gill [4] have compared algorithms such as principal component projections and the EM algorithm in discriminant analysis. More recently, Jurkowski and co-workers [5] compared different approaches to missing data in discrimination applied to medical problems based on Gower's distance.…”

Section: Introductionmentioning

confidence: 99%

Imputation techniques for incomplete data in quadratic discriminant analysis

Ounpraseuth

Moore

Young

2012

Journal of Statistical Computation and Simulation

View full text Add to dashboard Cite

We have compared the efficacy of five imputation algorithms readily available in SAS for the quadratic discriminant function. Here, we have generated several different parametric-configuration training data with missing data, including monotone missing-at-random observations, and used a Monte Carlo simulation to examine the expected probabilities of misclassification for the two-class quadratic statistical discrimination problem under five different imputation methods. Specifically, we have compared the efficacy of the complete observation-only method and the mean substitution, regression, predictive mean matching, propensity score, and Markov Chain Monte Carlo (MCMC) imputation methods. We found that the MCMC and propensity score multiple imputation approaches are, in general, superior to the other imputation methods for the configurations and training-sample sizes we considered.

show abstract

“…All the missing data are replacing with the value of mean available neighboring data [27]. Furthermore, other studies stated that mean substitution method is better and more accurate rather than eliminating the missing value with list wise and pairwise deletion method [28,29,30]. …”

Section: Data Cleaningmentioning

confidence: 99%

Monthly Analysis of Pm10 in Ambient Air of Klang Valley, Malaysia

Jamalani¹,

Abdullah²,

Azid³

et al. 2016

MJAS

View full text Add to dashboard Cite

The urbanization in Klang Valley, Peninsular Malaysia over the last decades has induce the atmospheric pollution's risk resulted to negative impact on the environment. The aims of this paper are to identify the spatial-temporal relationship of particulate matter (PM 10 ), to determine the characteristic of each location and to classify hierarchical of the location in relation to their impact on PM 10 concentration in Klang Valley. The Spearman correlation test indicate that there was strong significant relationship between all the locations (> 0.7; p < 0.001) and moderate relationship between Petaling Jaya-Kajang and Kajang-Shah Alam (< 0.7; p < 0.001). The principal component analysis (PCA) identifies all four locations have been affected by PM 10 which were determined as one of the pollutant that deteriorated the air quality. Cluster analysis (CA) has classified the PM 10 pattern into three (3) different classes; Class 1 (Klang), Class 2 (Petaling Jaya and Kajang) and Class 3 (Shah Alam) based on location. Further analysis of CA would be able to classify the PM 10 classes into groups depending on their dissimilarities characteristic. Thus, possible period of extreme air quality degradation could be identified. Therefore, statistical and envirometric techniques have proved the impact of the various location on increasing concentration of PM 10 . Keywords: particulate matter, Spearman correlation test, principal component analysis, cluster analysis AbstrakProses pembandaran di Lembah Klang, Semenanjung Malaysia sedekad lalu telah mendorong kepada risiko pencemaran atmosfera yang memberi impak negatif kepada alam sekitar. Kajian ini dilakukan bertujuan untuk mengenalpasti hubungkait antara ruang dan tempoh bagi partikel terampai (PM 10 ), menentukan ciri -ciri setiap lokasi dan menentukan pengkelasan hirarki lokasi berhubungan dengan impak kepekatan PM 10 di Lembah Klang. Ujian korelasi Spearman menunjukkan hubungkait yang kuat antara semua lokasi (> 0.7; p < 0.001) dan hubungan

show abstract

The Treatment of Missing Values in Discriminant Analysis—I. The Sampling Experiment

Cited by 34 publications

References 24 publications

The impact of four missing data techniques on validity estimates in human resource management

The impact of four missing data techniques on validity estimates in human resource management

Imputation techniques for incomplete data in quadratic discriminant analysis

Monthly Analysis of Pm10 in Ambient Air of Klang Valley, Malaysia

Contact Info

Product

Resources

About