隱藏分類分析 (Latent class analysis) 經常用以對無法直接觀測到變數的分析，例如在醫學的研究上，在真正疾病狀況 (即隱藏變數) 無法得知的情形下，我們會根據一些醫師的診斷結果 (即顯現變數) 進行分析，並藉此求得診斷的正確性。
傳統上隱藏變數分析最基本的假設是在給定隱藏分類的條件下，
各顯現變數之間彼此獨立，然而在實際的應用中，這樣的假設通常是不可行的。因此在這篇論文中，我們考慮的模式將納入給定隱藏分類下某些顯現變數之間的關聯性，我們對這樣的模式予以參數化，使得給定隱藏變數下，顯現變數的條件機率能以模式中參數的函數表示，而此條件機率通常是人們所感興趣的，並且我們利用邊際模式來說明存在於隱藏分類間顯現變數的相關性。
本研究亦將分類數由二個分類增加為三個分類，並介紹配適隱藏變數與顯現變數均為次序類別變數的模式。
在演算法方面利用改良於 EM 演算法的 EM 陡降演算法 (EM gradient algorithm) 來求得參數的最大概似估計，
並且估計這些估計量的準確性。本研究亦分別利用兩個實例來進行二分類與三分類隱藏分類邊際模式的配適。
研究結果發現，
1.獨立的隱藏分類模式不適切時，利用隱藏分類邊際模式可以得到一個合理(含有顯現變數的相關性)且易於解釋 (分類數相等) 的配適。
2.分類數增加後，矩陣規模增加的相當快速，雖然次序變數能減少參數個數，但就三分類與二分類比較，程式的執行時間增加的相當多。
基於實際的應用上，我們考慮具相關性且分類數多於二類的模式，然而在本研究著重於四個顯現變數的探討，四個以上的顯現變數是後續研究可以考慮的。

[1]中央研究院社會所， (1999). 台灣地區社會變遷基本調查資料庫。
http://140.109.196.2/sc1/scdown.htm (取得日期：2000,5,25)
[2]王光祿， (1995). 「離散資料同時模式之探討」，國立中央大學統計研究所未出版碩士論文。
[3]Agresti, A. (1990). Categorical Data Analysis. New York: Wiley.
[4]Alvord, W. G., Drummond, J. E., Arthur, L. O., Goedert, J. J., Levine, P. H., Murphy, E. L., Jr., Weiss, S. H. & Blattner, W. A. (1988).A Method for Predicting Individual HIV Infection Status in The Absence of Clinical Information. Aids Research and Human Retroviruses, 4, pp.295-304.
[5]Bartholomew, D. J. (1987). Latent Variable Models and Factor Analysis. Charles Griffin & Company Limited.
[6]Becker, M. P. & Yang, I. (1998). Latent Class Marginal Models for Cross-Classifications of counts. Sociological Methodology, pp.293-325.
[7]Becker, M. P., Yang, I. & Lange, K. (1997). EM Algorithm without Missing Data. Statistical Method in Medical Research, 6, pp.38-57.
[8]Clogg, C. C. (1979). Some Latent Structure Models for The Analysis of Likert-type Data. Science Research, 8, pp.287-301.
[9]Clogg, C. C. (1995). Latent Class Models. In Handbook of Statistical Modeling for the Social and Behavorial Sciences, G. Arminger, C. C. Clogg, & M. E. Sobel (eds), pp.311-359. New York: Plenum.
[10]Clogg, C. C. & Goodman, L. A. (1984). Latent Structure Analysis of A Set of Multidimensional Contingency Tables. Journal of the American Statistical Association, 79, pp.762-771.
[11]Espeland, M. A. & Handelman, S. L. (1989). Using Latent Class Models to Characterize and Access Relative Error in Discrete Measurements. Biometrics, 45, pp.587-599.
[12]Formann, A. K. (1978). The Latent Class Analysis of Polychotomous Data. Biom. J., Vol. 20, No. 718, pp.755-771.
[13]Formann, A. K. (1994). Measurement Errors in Caries Diagnosis: Some Further Latent Class Models. Biometrics, 50, pp.865-871.
[14]Goodman, L. A. (1974). Exploratory Latent structure analysis usint both identifiable and unidentifiable models. Biometrika, 61, pp.215-231.
[15]Haber, M. (1985). Maximum Likelihood Method for Linear and Log-Linear Models in Categorical Data. Computational Statistics & Data Analysis, 3, pp.1-10.
[16]Haberman, S. J. (1979). Analysis Qualitative Data, Volume 2, New Development. New York: Academic Press.
[17]Heinen, T. (1996). Latent Class and Discrete Latent Trait Models. Sage Publications.
[18]Lang, J. B. (1992). On Model fitting for Multivariate Polytomous Reponse Data. Unpublished dissertation, University of Folrida.
[19]Lang, J. B. (1993). Large Sample Behavior for A Broad Class of Multivariate Categorical Data Models. Unpublished Technical Report, University of Iowa.
[20]Lang, J. B. & Agresti, A. (1994). Simultaneously Modeling Joint and Marginal Distributions of Multivariate Categorical Responses. Journal of the American Statistical Association, Vol. 89, No. 426, pp.625-632.
[21]Lange, K. (1995). A Gradient Algorithm Locally Equivalent to The EM Algorithm. Journal of Royal Statistical Society, Series B 57, No. 2, pp.425-437.
[22]Lazarsfeld, P. F. (1950). Logical and Mathematical foundations of Latent structure Analysis. In Stouffer, S. A. et al.: Studies in Social Psychology in World War II, Vol. IV, Princeton Univ. Press.
[23]Lazarsfeld, P. F. & Henry, N. W. (1968). Latent Structure Analysis. Boston: Houghton Mifflin Co.
[24]McHugh, R. B. (1956). Efficient Estimation and Local Identification in Latent Class Analysis. Psycometrika, Vol. 21, No. 4, pp.331-347.
[25]Meng, X. & Rubin, D. B. (1991). Using EM to Obtain Asymptotic Variance-Covariance Matrices: The SEM Algorithm. Journal of the American Statistical Association, Vol. 86, No. 416, pp.899-909.
[26]Poulsen, C. S. (1982). Latent Structure Analysis with Choice Modeling Application. doctoral dissertation, The Arhus School of Business Administration and Economic.
[27]Qu, Y. & Hadgu, A. (1998). A Model for Evaluating Sensitivity and Specificity for Correlated Diagnostic Tests in Efficacy Studies With an Imperfect Referenct Test. Journal of the American Statistical Association, Vol. 93, No. 443, pp.920-928.
[28]Qu, Y., Tan, M. & Kutner, M. H. (1996). Random Effects Models in Latent Class analysis for Evaluating Accuracy of Diagnostic Tests. Biometrics, 52, pp.797-810.
[29]Ramaswamy, V., Chatterjee, R. & Cohen, S. H. (1999). Reply to "A note on Ramaswamy, Chatterjee, and Cohen''s Latent Joint Segmentation Models". Journal of Marketing Research Vol. XXXVI, pp.115-119.
[30]Uebersax, J. S. & Grove, W. M. (1990). Latent Class analysis of Diagnostic Agreement. Statistics in Medicine, 9, pp.559-572.
[31]Yang, I. & Becker, M. P. (1997). Latent Variable Modeling of Diagnostic Accuracy. Biometrics, 53, pp.948-958.