Statistical analysis of maintenance growth curves

Abstract

Selecting the number of pace categories using cross validation Conclusion 3 ESTIMATION AND PREDICTION OF LABOR HOURS USING LOCALLY WEIGHTED REGRESSION 23 Application to the truck data 27 Computation of rh{t) 33 Diagnostics 34 Bootstrap estimates for standaxd error of m(f) and Ay{ta,U) 35 Incorporating pace into the loess analysis Conclusion 43 4 GROWTH CURVE ESTIMATION AND PREDICTION USING LINEAR MIXED MODELS Using the model for prediction of future values 58 Interpretation of 7,and y-p 60 Choosing a model via cross validation 61 Checking model assumptions 62 Incorporating pace into the mixed model analysis 64 Discussion and conclusion 67 5 A COMPARISON OF METHODS USING 1994 DATA 69 c triangular factorization of the covaxiance matrix, c NR, maximum number of nonlineax parameters, c MODEL(1)+MODEL(2)+MODEL(3)+NOD c NCV, number of covaariates or grouping variables, constant for each c subj ect. c NTVCV, number of covauriates that vary within each subject's data c NLP, number of linear pairameters c ND must be at least NLP+1 c NS, number of subjects c NMAX, msLximum number of observations c