v0.2.0

.gitignore

@@ -1,6 +1,9 @@
 src/temp.jl
 test/precompile/
 test/devtest.jl
+test/dstest.jl
+test/snoopy.jl
 examples/
 docs/build/
 docs/img/
+etc/

Project.toml

@@ -1,7 +1,7 @@
 name = "Metida"
 uuid = "a1dec852-9fe5-11e9-361f-8d9fde67cfa2"
 authors = ["Vladimir Arnautov <[email protected]>"]
-version = "0.1.14"
+version = "0.2.0"
 
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
@@ -29,7 +29,8 @@ julia = "1.1, 1.2, 1.3, 1.4, 1.5, 1.6"
 CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
+StatsModels = "3eaba693-59b7-5ba5-a881-562e759f1c8d"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "CSV", "DataFrames", "StatsBase"]
+test = ["Test", "CSV", "DataFrames", "StatsBase", "StatsModels"]

change.log

@@ -1,3 +1,11 @@
+- 0.2.0
+  * many changes in VarianceType, covariance structure
+  * code cleaning
+  * change methods names
+  * rmatrix / gmatrix
+  * StatsBase methods
+  * test and other
+
 - 0.1.13
 
   * output fix

docs/src/api.md

@@ -49,3 +49,13 @@ Metida.CompoundSymmetry
 ```@docs
 Metida.HeterogeneousCompoundSymmetry
 ```
+
+### Metida.gmatrix
+```@docs
+Metida.gmatrix
+```
+
+### Metida.rmatrix
+```@docs
+Metida.rmatrix
+```

docs/src/ref.md

@@ -3,18 +3,74 @@
 ### Sweep algorithm
 
 * Sweep based on [SweepOperator.jl](https://github.com/joshday/SweepOperator.jl). Thanks to @joshday and @Hua-Zhou.
+
 * http://hua-zhou.github.io/teaching/biostatm280-2019spring/slides/12-sweep/sweep.html
 
 More:
 
-*  Section 7.4-7.6 of Numerical Analysis for Statisticians by Kenneth Lange (2010).
+* Section 7.4-7.6 of Numerical Analysis for Statisticians by Kenneth Lange (2010).
+
 * The paper A tutorial on the SWEEP operator by James H. Goodnight (1979).
 
 
-### REML
+### REML & Parameter estimation
+
+* Henderson, C. R., et al. “The Estimation of Environmental and Genetic Trends from Records Subject to Culling.” Biometrics, vol. 15, no. 2, 1959, pp. 192–218. JSTOR, www.jstor.org/stable/2527669.
+
+* Lindstrom & J.; Bates, M. (1988). Newton—Raphson and EM Algorithms for Linear Mixed-Effects Models for Repeated-Measures Data. Journal of the American Statistical Association. 83. 1014. 10.1080/01621459.1988.10478693.
 
+* Gurka, Matthew. (2006). Selecting the Best Linear Mixed Model under REML. The American Statistician. 60. 19-26. 10.1198/000313006X90396.
+
+* Sang Hong Lee, Julius H.J. van der Werf. An efficient variance component approach implementing an average information REML suitable for combined LD and linkage mapping with a general complex pedigree. Genetics Selection Evolution, BioMed Central, 2006, 38 (1), pp.25-43. ⟨hal-00894558⟩
+
+* F.N. Gumedze, T.T. Dunne, Parameter estimation and inference in the linear mixed model, Linear Algebra and its Applications, Volume 435, Issue 8, 2011, Pages 1920-1944, ISSN 0024-3795, https://doi.org/10.1016/j.laa.2011.04.015. (http://www.sciencedirect.com/science/article/pii/S002437951100320X)
 
 ### AI algorithm
 
+* D.L. Johnson, R. Thompson, Restricted Maximum Likelihood Estimation of Variance Components for Univariate Animal Models Using Sparse Matrix Techniques and Average Information, Journal of Dairy Science, Volume 78, Issue 2, 1995, Pages 449-456, ISSN 0022-0302, https://doi.org/10.3168/jds.S0022-0302(95)76654-1. (http://www.sciencedirect.com/science/article/pii/S0022030295766541)
+
+* Mishchenko, Kateryna & Holmgren, Sverker & Rönnegård, Lars. (2007). Newton-type Methods for REML Estimation in Genetic Analysis of Quantitative Traits. Journal of Computational Methods in Science and Engineering. 8. 10.3233/JCM-2008-81-203.
+
+* Matilainen K, Mäntysaari EA, Lidauer MH, Strandén I, Thompson R. Employing a Monte Carlo algorithm in Newton-type methods for restricted maximum likelihood estimation of genetic parameters. PLoS One. 2013;8(12):e80821. Published 2013 Dec 10. doi:10.1371/journal.pone.0080821
 
 ### Covariance structures
+
+* Wolfinger, Russ. (1993). Covariance structure selection in general mixed models. Communications in Statistics-simulation and Computation - COMMUN STATIST-SIMULAT COMPUT. 22. 1079-1106. 10.1080/03610919308813143.
+
+* Littell, Ramon & Pendergast, Jane & Natarajan, Ranjini. (2000). Modelling covariance structure in the analysis of repeated measures data. Statistics in Medicine. 19. 1793-1819. 10.1002/1097-0258(20000715)19:13%3C1793::AID-SIM482%3E3.0.CO;2-Q.
+
+* QUINTAL, SILVANA SILVA RED, VIANA, ALEXANDRE PIO, CAMPOS, BIANCA MACHADO, VIVAS, MARCELO, & AMARAL JÚNIOR, ANTONIO TEIXEIRA DO. (2017). ANALYSIS OF STRUCTURES OF COVARIANCE AND REPEATABILITY IN GUAVA SEGREGANTING POPULATION. Revista Caatinga, 30(4), 885-891. https://doi.org/10.1590/1983-21252017v30n408rc
+
+* McNeish, D., Harring, J. Covariance pattern mixture models: Eliminating random effects to improve convergence and performance. Behav Res 52, 947–979 (2020). https://doi.org/10.3758/s13428-019-01292-4
+
+### And more
+
+* Laird, Nan M., and James H. Ware. “Random-Effects Models for Longitudinal Data.” Biometrics, vol. 38, no. 4, 1982, pp. 963–974. JSTOR, www.jstor.org/stable/2529876.
+
+* Giesbrecht, F. G., and Burns, J. C. (1985), "Two-Stage Analysis Based on a Mixed Model: Large-sample Asymptotic Theory and Small-Sample Simulation Results," Biometrics, 41, 853-862.
+
+* Jennrich, R., & Schluchter, M. (1986). Unbalanced Repeated-Measures Models with Structured Covariance Matrices. Biometrics, 42(4), 805-820. doi:10.2307/2530695
+
+* Fletcher, Roger (1987), Practical methods of optimization (2nd ed.), New York: John Wiley & Sons, ISBN 978-0-471-91547-8
+
+* Wolfinger et al., (1994) Computing gaussian likelihoods and their derivatives for general linear mixed models doi: 10.1137/0915079
+
+* Hrong-Tai Fai & Cornelius (1996) Approximate F-tests of multiple degree of freedom hypotheses in generalized least squares analyses of unbalanced split-plot experiments, Journal of Statistical Computation and Simulation, 54:4, 363-378, DOI: 10.1080/00949659608811740
+
+* Schaalje GB, McBride JB, Fellingham GW. Adequacy of approximations to distributions of test statistics in complex mixed linear models. J Agric Biol Environ Stat. 2002;7:512–24.
+
+* Wright, Stephen, and Jorge Nocedal (2006) "Numerical optimization." Springer
+
+* Van Peer, A. (2010), Variability and Impact on Design of Bioequivalence Studies. Basic & Clinical Pharmacology & Toxicology, 106: 146-153. doi:10.1111/j.1742-7843.2009.00485.x
+
+### Julia packages
+
+* Revels, Jarrett & Lubin, Miles & Papamarkou, Theodore. (2016). Forward-Mode Automatic Differentiation in Julia.
+
+* Mogensen et al., (2018). Optim: A mathematical optimization package for Julia. Journal of Open Source Software, 3(24), 615,doi: 10.21105/joss.00615
+
+### CuSolver & CuBLAS
+
+* https://docs.nvidia.com/cuda/cusolver/index.html
+
+* https://docs.nvidia.com/cuda/cublas/index.htm

docs/src/validation.md

@@ -1,6 +1,81 @@
 ## Validation
 
+### sleepstudy.csv
 
+| Model  |  REML  Metida | REML SPSS|
+|--------|--------|-------|
+| 1 | 1729.4925602367025 | 1729.492560 |
+| 2 | 1904.3265170722132 | 1662.172084 |
+| 3 | 1772.0953251997046 | 1772.095 |
+| 4 | 1730.1895427398322 | 1730.189543 |
+
+
+#### Model 1
+
+```
+lmm = Metida.LMM(@formula(Reaction~Days), df;
+  random = Metida.VarEffect(Metida.SI),
+  subject = :Subject
+  )
+  Metida.fit!(lmm)
+```
+
+```
+```
+#### Model 2
+
+```
+lmm = Metida.LMM(@formula(Reaction~1), df;
+  random = Metida.VarEffect(Metida.@covstr(Days), Metida.CS),
+  subject = :Subject
+  )
+  Metida.fit!(lmm)
+```
+
+```
+```
+
+#### Model 3
+
+```
+lmm = Metida.LMM(@formula(Reaction~1), df;
+  random = Metida.VarEffect(Metida.@covstr(Days), Metida.CSH, subj = :Subject)
+  )
+  Metida.fit!(lmm)
+```
+
+```
+```
+
+#### Model 4
+
+```
+lmm = Metida.LMM(@formula(Reaction~1), df;
+  random = Metida.VarEffect(Metida.@covstr(Days), Metida.ARH, subj = :Subject)
+  )
+  Metida.fit!(lmm)
+```
+
+```
+```
+
+### Pastes.csv
+
+| Model  |  REML  Metida | REML SPSS|
+|--------|--------|-------|
+| 5 | 246.99074585348623 | 246.990746 |
+
+#### Model 5
+
+```
+lmm = Metida.LMM(@formula(strength~1), df;
+random = [Metida.VarEffect(Metida.SI, subj = :batch), Metida.VarEffect(Metida.SI, subj = [:batch,  :cask])]
+)
+Metida.fit!(lmm)
+```
+
+```
+```
 
 ### Reference dataset
 
@@ -20,4 +95,4 @@ Anglin, P.M. and R. Gencay (1996) “Semiparametric estimation of a hedonic pric
 Anglin, P., and Gencay, R. (1996). Semiparametric Estimation of a Hedonic Price Function. Journal of Applied Econometrics, 11, 633–648.
 
 Verbeek, M. (2004). A Guide to Modern Econometrics, 2nd ed. Chichester, UK: John Wiley.
-O'Brien, R. G., and Kaiser, M. K. (1985) MANOVA method for analyzing repeated measures designs: An extensive primer. Psychological Bulletin 97, 316–333, Table 7. 
+O'Brien, R. G., and Kaiser, M. K. (1985) MANOVA method for analyzing repeated measures designs: An extensive primer. Psychological Bulletin 97, 316–333, Table 7.

src/Metida.jl

@@ -25,7 +25,6 @@ fit!, LMM, VarEffect, theta, logreml, m2logreml, thetalength
 
 include("abstracttype.jl")
 include("sweep.jl")
-include("utils.jl")
 include("varstruct.jl")
 include("gmat.jl")
 include("rmat.jl")
@@ -40,6 +39,7 @@ include("fit.jl")
 include("showutils.jl")
 include("statsbase.jl")
 include("initg.jl")
+include("utils.jl")
 
 function __init__()
 
@@ -48,16 +48,24 @@ end
 function _precompile_()
     ccall(:jl_generating_output, Cint, ()) == 1 || return nothing
     isdefined(Metida, Symbol("#@covstr")) && precompile(Tuple{getfield(Metida, Symbol("#@covstr")), LineNumberNode, Module, Int})
-    precompile(Tuple{typeof(Metida.ffx), Int64})
-    precompile(Tuple{typeof(Metida.ffxpone), Int64})
 
     precompile(Tuple{typeof(Metida.reml_sweep_β), Metida.LMM{Float64}, Array{Float64, 1}})
     precompile(Tuple{typeof(Metida.reml_sweep_β_b), Metida.LMM{Float64}, Array{Float64, 1}})
 
     precompile(Tuple{typeof(Metida.rholinkpsigmoid), Float64})
+    precompile(Tuple{typeof(Metida.rholinkpsigmoidr), Float64})
+
+    precompile(Tuple{typeof(Metida.varlinkvecapply!),  Array{Float64, 1},  Array{Symbol, 1}})
+    precompile(Tuple{typeof(Metida.varlinkrvecapply!),  Array{Float64, 1},  Array{Symbol, 1}})
+    precompile(Tuple{typeof(Metida.varlinkvecapply),  Array{Float64, 1},  Array{Symbol, 1}})
 
-    precompile(Tuple{typeof(Metida.varlinkvecapply!), Array{Float64, 1}, Array{Function, 1}})
     precompile(Tuple{typeof(Metida.vlink), Float64})
+    precompile(Tuple{typeof(Metida.vlinkr), Float64})
+
+    precompile(Tuple{typeof(Metida.initvar),  Array{Float64, 1},  Array{Float64, 2}})
+
+    precompile(Tuple{typeof(Metida.intersectsubj),  Array{VarEffect, 1},  VarEffect})
+
 end
 _precompile_()
 #include(".jl")