Linear Discriminant Analysis Versus Logistic Regression: A Comparison of Classification Errors in the Two-Group Case

Abstract Classification studies are important for practitioners who need to identify individuals for specialized treatment or intervention. When interventions are irreversible or misclassifications are costly, information about the proficiency of different classification procedures becomes invaluable. This study furnishes information about the relative accuracy of two widely used classification procedures, linear discriminant analysis and logistic regression, under various commonly encountered and interacting conditions. Monte Carlo simulation was used to manipulate four factors under multivariate normality: equality of covariance matrices, degree of group separation, sample size, and prior probabilities. Three criterion measures were employed: total, small-group, and large-group classification error. Interactions of these between factors with two within factors, cut-score and method of classification, were of primary interest.

[1]  E. J. Kelly,et al.  Distinguishing between Conduct Problem and Emotionally Disturbed Students in Secondary School: A Five-Instrument Discriminant Analysis , 1991, Psychological reports.

[2]  B. Fowers,et al.  Enrich marital inventory: a discriminant validity and cross-validation assessment. , 1989, Journal of marital and family therapy.

[3]  G. Glass,et al.  Statistical methods in education and psychology , 1970 .

[4]  A. Baron,et al.  Misclassification among methods used for multiple group discrimination--the effects of distributional properties. , 1991, Statistics in medicine.

[5]  John C. Lind,et al.  Two-sample T–2 procedure and the assumption of homogeneous covariance matrices. , 1979 .

[6]  S. J. Press,et al.  Choosing between Logistic Regression and Discriminant Analysis , 1978 .

[7]  John D. Morris,et al.  Predictive Discriminant Analysis Versus Logistic Regression in Two-Group Classification Problems. , 1996 .

[8]  P. Meehl,et al.  Antecedent probability and the efficiency of psychometric signs, patterns, or cutting scores. , 1955, Psychological bulletin.

[9]  Larry J. Kortering,et al.  The Identification of High-School Dropouts Identified as Learning Disabled: Evaluating the Utility of a Discriminant Analysis Function , 1992, Exceptional children.

[10]  D. Benson,et al.  Behavioral Phenomenology in Alzheimer's Disease, Frontotemporal Dementia, and Late-Life Depression: A Retrospective Analysis , 1997, Journal of geriatric psychiatry and neurology.

[11]  Alan J. Gumm,et al.  The Development of a Model and Assessment Instrument of Choral Music Teaching Styles , 1993 .

[12]  Constantine Loucopoulos,et al.  Three-group classification with unequal misclassification costs: a mathematical programming approach , 2001 .

[13]  K. Marjoribanks.  Environmental and Individual Influences on Australian Students' Likelihood of Attending University , 1998 .

[14]  C. Heckler.  Applied Discriminant Analysis , 1995 .

[15]  Bill C. Hardgrave,et al.  Predicting Graduate Student Success in an MBA Program: Regression Versus Classification , 1995 .

[16]  J. Gfroerer,et al.  Substance use in the US college-age population: differences according to educational status and living arrangement. , 1997, American journal of public health.

[17]  David W. Hosmer,et al.  Applied Logistic Regression , 1991 .

[18]  G. Glass,et al.  Statistical methods in education and psychology, 3rd ed. , 1996 .

[19]  Charles E. Heckler,et al.  Applied Multivariate Statistical Analysis , 2005, Technometrics.

[20]  Jane Corville-Smith,et al.  Distinguishing Absentee Students from Regular Attenders: The Combined Influence of Personal, Family, and School Factors , 1998 .

[21]  Kern W. Dickman,et al.  Sample and population score matrices and sample correlation matrices from an arbitrary population correlation matrix , 1962 .

[22]  E. J. Kelly,et al.  Distinguishing between Conduct-Problem and Emotionally Disturbed Students in Elementary School: A Five-Instrument Discriminant Analysis , 1992, Psychological reports.

[23]  J. Modestin,et al.  A Study of Clinical Suicide , 1988, Journal of Nervous and Mental Disease.

[24]  E. Abemayor,et al.  Smoking withdrawal and relapse in head and neck cancer patients , 1999, Head & neck.

[25]  L. Leemis.  Applied Linear Regression Models , 1991 .

[26]  D. Lezotte,et al.  Optimal Gestational Age and Birth-weight Cutoffs to Predict Neonatal Morbidity , 2000, Medical decision making : an international journal of the Society for Medical Decision Making.

[27]  A. Bello,et al.  Misclassification among methods used for multiple group discrimination--the effects of distributional properties. , 1992, Statistics in medicine.

[28]  S. Cadzow,et al.  Stressed parents with infants: reassessing physical abuse risk factors. , 1999, Child abuse & neglect.

[29]  T. Blum,et al.  Organizational Characteristics and Employer Responses to Employee Substance Abuse , 1998 .

[30]  B. Efron.  The Efficiency of Logistic Regression Compared to Normal Discriminant Analysis , 1975 .

[31]  R. Breteler,et al.  The discriminant validity of the Eating Disorder Inventory--2. , 1997, The British journal of clinical psychology.

[32]  Mark R. Wade,et al.  Construction and Assessment of Classification Rules , 1999, Technometrics.

[33]  J. Neter,et al.  Applied Linear Regression Models , 1983 .

[34]  C. Lyketsos,et al.  Relationship of cognitive and functional impairment to depressive features in Alzheimer's disease and other dementias. , 1998, The Journal of neuropsychiatry and clinical neurosciences.

[35]  Guido Rodriguez,et al.  Quantitative Electroencephalography and Regional Cerebral Blood Flow: Discriminant Analysis between Alzheimer’s Patients and Healthy Controls , 1998, Dementia and Geriatric Cognitive Disorders.

[36]  Xitao Fan,et al.  Comparing Linear Discriminant Function with Logistic Regression for the Two-Group Classification Problem. , 1999 .