[[Ｒ備忘録 - 記事一覧]] !!!RSNNS *投稿者： みゅ *カテゴリ： なし *優先度： 普通 *状態： 完了 *日時： 2011年04月20日 22時42分18秒 //{{bugstate}} !!内容 *メモ *http://cran.r-project.org/web/packages/RSNNS/RSNNS.pdf *http://www.aso.ecei.tohoku.ac.jp/publication_data/346.pdf *http://www.ra.cs.uni-tuebingen.de/SNNS/ *[神経回路網による自己連想記憶モデルに関する実験|http://150.55.136.242/lecture/jiken/1ps_semester.pdf] *[5.1 連想記憶|http://www.gifu-nct.ac.jp/elec/deguchi/sotsuron/owaki/node16.html] !!スクリプト library(nnet) library(RSNNS) data(snnsData) patterns <- snnsData$art1_letters.pat par(mfrow=c(3,3)) for (i in 1:9) plotActMap(matrixToActMapList(patterns,7)[[i]]) res <- nnet(patterns, y=patterns, skip=T, size=35, MaxNWts=20000, maxit=1000, lineout=F) par(mfrow=c(3,3)) for (i in 1:9) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) for (i in 10:18) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) for (i in 19:26) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) testData <- rbind( as.vector(t(rbind(c(0,0,0,0,1), c(0,0,0,1,0), c(0,0,0,1,0), c(0,0,1,0,0), c(0,1,0,0,0), c(0,1,0,0,0), c(1,0,0,0,0)))), as.vector(t(rbind(c(1,0,0,0,0), c(1,0,0,0,0), c(0,1,0,0,0), c(0,0,1,0,0), c(0,0,0,1,0), c(0,0,0,0,1), c(0,0,0,0,1)))) ) par(mfrow=c(2,2)) plotActMap(matrixToActMapList(testData,7)[[1]]) plotActMap(matrixToActMapList(testData,7)[[2]]) pred <- predict(res, testData) plotActMap(matrixToActMapList(pred,7)[[1]]) plotActMap(matrixToActMapList(pred,7)[[2]]) par(mfrow=c(2,2)) plotActMap(matrixToActMapList(testData,7)[[1]]) plotActMap(matrixToActMapList(testData,7)[[2]]) pred <- predict(res, pred) plotActMap(matrixToActMapList(pred,7)[[1]]) plotActMap(matrixToActMapList(pred,7)[[2]]) library(genalg) func <- function(param){ res$wts <- param #sum(abs(patterns - predict(res,patterns)) %*% rep(1,35)) sum((patterns - predict(res,patterns))^2 %*% rep(1,35)) } func(runif(length(res$wts))) monFunc <- function(obj){ res$wts <- obj$population[1,] pred <- predict(res, patterns) par(mfrow=c(3,3)) plotActMap(matrixToActMapList(patterns,7)[[1]]) plotActMap(matrixToActMapList(patterns,7)[[2]]) plotActMap(matrixToActMapList(patterns,7)[[3]]) plotActMap(matrixToActMapList(pred,7)[[1]]) plotActMap(matrixToActMapList(pred,7)[[2]]) plotActMap(matrixToActMapList(pred,7)[[3]]) plot(obj$best, ty="l") lines(obj$mean, col="red") #for (i in 1:9) plotActMap(matrixToActMapList(pred,7)[[i]]) } monFunc(res_ga) res_ga <- rbga(stringMin=rep(-1,length(res$wts)), stringMax=rep(1,length(res$wts)), evalFunc=func, monitorFunc=monFunc, verbose=T, iters=200, suggestions=rbind(res$wts,res$wts)) res_ga$population[1,] res <- nnet(patterns, y=patterns, skip=T, size=35, MaxNWts=20000, maxit=1000, lineout=F, Wts=res_ga$population[1,]) par(mfrow=c(3,3)) for (i in 1:9) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) for (i in 10:18) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) for (i in 19:26) plotActMap(matrixToActMapList(res$fitted,7)[[i]]) res_ga <- rbga(stringMin=rep(-1,length(res$wts)), stringMax=rep(1,length(res$wts)), evalFunc=func, monitorFunc=monFunc, verbose=T, iters=200, suggestions=rbind(res$wts,res$wts)) # パターンをひっくり返す patterns2 <- 1-patterns par(mfrow=c(3,3)) for (i in 1:9) plotActMap(matrixToActMapList(patterns2,7)[[i]]) pred <- predict(res, patterns2) for (i in 1:9) plotActMap(matrixToActMapList(pred,7)[[i]]) *art1 library(RSNNS) data(snnsData) patterns <- snnsData$art1_letters.pat inputMaps <- matrixToActMapList(patterns, nrow=7) par(mfrow=c(3,3)) for (i in 1:9) plotActMap(inputMaps[[i]]) model <- art1(patterns, dimX=7, dimY=5) encodeClassLabels(model$fitted.values) > ls(envir=model$snnsObject@variables) [1] "serialization" "snnsCLibPointer" predict(model, newdata=patterns) encodeClassLabels(predict(model, newdata=patterns)) summary(model) *art2 library(RSNNS) data(snnsData) patterns <- snnsData$art2_tetra_med.pat model <- art2(patterns, f2Units=5, learnFuncParams=c(0.99, 20, 20, 0.1, 0), updateFuncParams=c(0.99, 20, 20, 0.1, 0)) model testPatterns <- snnsData$art2_tetra_med.pat predictions <- predict(model, testPatterns) library(scatterplot3d) par(mfrow=c(1,2)) scatterplot3d(model$fitted.values[,1:3]) scatterplot3d(predictions[,1:3]) model *dlvq library(RSNNS) demo(dlvq_ziff) data(snnsData) dataset <- snnsData$dlvq_ziff_100.pat inputs <- dataset[,inputColumns(dataset)] outputs <- dataset[,outputColumns(dataset)] model <- dlvq(inputs, outputs) mean(fitted(model) - outputs) confusionMatrix(outputs, fitted(model)) par(mfrow=c(2,2)) plotActMap(matrixToActMapList(inputs, 16)[[1]]) plotActMap(matrixToActMapList(inputs, 16)[[2]]) plotActMap(matrixToActMapList(inputs, 16)[[3]]) plotActMap(matrixToActMapList(inputs, 16)[[4]]) plotActMap(matrixToActMapList(inputs, 16)[[5]]) plotActMap(matrixToActMapList(inputs, 16)[[50]]) plotActMap(matrixToActMapList(inputs, 16)[[15]]) plotActMap(matrixToActMapList(inputs, 16)[[2]]) plotActMap(matrixToActMapList(inputs, 16)[[7]]) plotActMap(matrixToActMapList(inputs, 16)[[13]]) patterns <- dataset[,inputColumns(dataset)][c(1,2,3,4,5,6,8,10,11,12),] plotActMap(matrixToActMapList(patterns,16)[[1]]) *iris > library(RSNNS) > seed <- 2 > set.seed(seed) > setSnnsRSeedValue(seed) NULL > data(iris) > #shuffle the vector > iris <- iris[sample(1:nrow(iris),length(1:nrow(iris))),1:ncol(iris)] > irisValues <- iris[,1:4] > irisTargets <- decodeClassLabels(iris[,5]) > #irisTargets <- decodeClassLabels(iris[,5], valTrue=0.9, valFalse=0.1) > > iris <- splitForTrainingAndTest(irisValues, irisTargets, ratio=0.15) > #normalize data > iris <- normTrainingAndTestSet(iris) > #model <- mlp(iris$inputsTrain, iris$targetsTrain, size=5, learnFunc="Quickprop", learnFuncParams=c(0.1, 2.0, 0.0001, 0.1), > # maxit=50, inputsTest=iris$inputsTest, targetsTest=iris$targetsTest) > > model <- mlp(iris$inputsTrain, iris$targetsTrain, size=5, learnFunc="BackpropBatch", learnFuncParams=c(10, 0.1), + maxit=100, inputsTest=iris$inputsTest, targetsTest=iris$targetsTest) > #model <- mlp(iris$inputsTrain, iris$targetsTrain, size=5, learnFunc="SCG", learnFuncParams=c(0, 0, 0, 0), > # maxit=30, inputsTest=iris$inputsTest, targetsTest=iris$targetsTest) > > > #model <- rbfDDA(iris$inputsTrain, iris$targetsTrain) > > #model <- elman(iris$inputsTrain, iris$targetsTrain, size=5, learnFuncParams=c(0.1), maxit=100, inputsTest=iris$inputsTest, targetsTest=iris$targetsTest) > > #model <- rbf(iris$inputsTrain, iris$targetsTrain, size=40, maxit=200, initFuncParams=c(-4, 4, 0.0, 0.2, 0.04), > # learnFuncParams=c(1e-3, 0, 1e-3, 0.1, 0.8), linOut=FALSE) > > #model <- rbf(iris$inputsTrain, iris$targetsTrain, size=40, maxit=600, initFuncParams=c(0, 1, 0.0, 0.2, 0.04), > # learnFuncParams=c(1e-5, 0, 1e-5, 0.1, 0.8), linOut=TRUE) > > ##experimental..: > ##model <- rbf(iris$inputsTrain, iris$targetsTrain, size=20, maxit=50, initFunc="RBF_Weights_Kohonen", > ## initFuncParams=c(50, 0.4, 0), learnFuncParams=c(0.01, 0, 0.01, 0.1, 0.8)) > > #summary(model) > > par(mfrow=c(2,2)) > plotIterativeError(model) 次の図を見るためにはキーを押して下さい: > predictions <- predict(model,iris$inputsTest) > plotRegressionError(predictions[,2], iris$targetsTest[,2]) > confusionMatrix(iris$targetsTrain,fitted.values(model)) predictions targets 1 2 3 1 41 0 0 2 0 45 1 3 0 0 40 > confusionMatrix(iris$targetsTest,predictions) predictions targets 1 2 3 1 9 0 0 2 0 3 1 3 0 1 9 > plotROC(fitted.values(model)[,2], iris$targetsTrain[,2]) > plotROC(predictions[,2], iris$targetsTest[,2]) > #confusion matrix with 402040-method > confusionMatrix(iris$targetsTrain, encodeClassLabels(fitted.values(model),method="402040", l=0.4, h=0.6)) predictions targets 0 1 2 3 1 0 41 0 0 2 2 0 43 1 3 2 0 0 38 > model Class: mlp->rsnns Number of inputs: 4 Number of outputs: 3 Maximal iterations: 100 Initialization function: Randomize_Weights Initialization function parameters: -0.3 0.3 Learning function: BackpropBatch Learning function parameters: 10 0.1 Update function:Topological_Order Update function parameters: 0 Patterns are shuffled internally: TRUE Compute error in every iteration: TRUE Architecture Parameters: $size [1] 5 All members of model: [1] "nInputs" "maxit" "initFunc" "initFuncParams" "learnFunc" "learnFuncParams" [7] "updateFunc" "updateFuncParams" "shufflePatterns" "computeIterativeError" "snnsObject" "archParams" [13] "IterativeFitError" "IterativeTestError" "fitted.values" "fittedTestValues" "nOutputs" > weightMatrix(model) Input_1 Input_2 Input_3 Input_4 Hidden_2_1 Hidden_2_2 Hidden_2_3 Hidden_2_4 Hidden_2_5 Output_1 Output_2 Output_3 Input_1 0 0 0 0 -0.3508658 0.2684588 -0.5913861 -0.5612198 0.3195967 0.00000000 0.000000 0.000000 Input_2 0 0 0 0 -0.5082296 -0.9437067 1.2823328 0.9810203 0.2950290 0.00000000 0.000000 0.000000 Input_3 0 0 0 0 2.2801065 1.2012331 -1.8615422 -1.2803886 -1.9130263 0.00000000 0.000000 0.000000 Input_4 0 0 0 0 2.7568083 1.5704848 -1.6876845 -1.3023156 -2.6831374 0.00000000 0.000000 0.000000 Hidden_2_1 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -1.40362704 -3.796962 3.113594 Hidden_2_2 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -2.07435131 1.398071 1.325584 Hidden_2_3 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 2.11041474 -3.148410 -2.083688 Hidden_2_4 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 1.70966542 -1.791131 -2.028595 Hidden_2_5 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -0.04861032 3.213499 -3.448702 Output_1 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 Output_2 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 Output_3 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 > summary(model) SNNS network definition file V1.4-3D generated at Mon Sep 3 14:54:38 2012 network name : RSNNS_untitled source files : no. of units : 12 no. of connections : 35 no. of unit types : 0 no. of site types : 0 learning function : BackpropBatch update function : Topological_Order unit default section : act | bias | st | subnet | layer | act func | out func ---------|----------|----|--------|-------|--------------|------------- 0.00000 | 0.00000 | i | 0 | 1 | Act_Logistic | Out_Identity ---------|----------|----|--------|-------|--------------|------------- unit definition section : no. | typeName | unitName | act | bias | st | position | act func | out func | sites ----|----------|------------|----------|----------|----|----------|--------------|----------|------- 1 | | Input_1 | 0.81788 | 0.12059 | i | 1,0,0 | Act_Identity | | 2 | | Input_2 | -0.09058 | 0.18581 | i | 2,0,0 | Act_Identity | | 3 | | Input_3 | 0.82358 | -0.24672 | i | 3,0,0 | Act_Identity | | 4 | | Input_4 | 1.07735 | -0.22711 | i | 4,0,0 | Act_Identity | | 5 | | Hidden_2_1 | 0.82712 | -3.04162 | h | 1,2,0 ||| 6 | | Hidden_2_2 | 0.97982 | 0.89637 | h | 2,2,0 ||| 7 | | Hidden_2_3 | 0.00414 | -1.53274 | h | 3,2,0 ||| 8 | | Hidden_2_4 | 0.01841 | -0.97075 | h | 4,2,0 ||| 9 | | Hidden_2_5 | 0.17175 | 2.65829 | h | 5,2,0 ||| 10 | | Output_1 | 0.01332 | -1.14370 | o | 1,4,0 ||| 11 | | Output_2 | 0.12910 | -0.64419 | o | 2,4,0 ||| 12 | | Output_3 | 0.88143 | -1.22978 | o | 3,4,0 ||| ----|----------|------------|----------|----------|----|----------|--------------|----------|------- connection definition section : target | site | source:weight -------|------|--------------------------------------------------------------------------------------------------------------------- 5 | | 4: 2.75681, 3: 2.28011, 2:-0.50823, 1:-0.35087 6 | | 4: 1.57048, 3: 1.20123, 2:-0.94371, 1: 0.26846 7 | | 4:-1.68768, 3:-1.86154, 2: 1.28233, 1:-0.59139 8 | | 4:-1.30232, 3:-1.28039, 2: 0.98102, 1:-0.56122 9 | | 4:-2.68314, 3:-1.91303, 2: 0.29503, 1: 0.31960 10 | | 9:-0.04861, 8: 1.70967, 7: 2.11041, 6:-2.07435, 5:-1.40363 11 | | 9: 3.21350, 8:-1.79113, 7:-3.14841, 6: 1.39807, 5:-3.79696 12 | | 9:-3.44870, 8:-2.02859, 7:-2.08369, 6: 1.32558, 5: 3.11359 -------|------|--------------------------------------------------------------------------------------------------------------------- > extractNetInfo(model) $infoHeader name value 1 no. of units 12 2 no. of connections 35 3 no. of unit types 0 4 no. of site types 0 5 learning function BackpropBatch 6 update function Topological_Order $unitDefinitions unitNo unitName unitAct unitBias type posX posY posZ actFunc outFunc sites 1 1 Input_1 0.817876339 0.1205858 UNIT_INPUT 1 0 0 Act_Identity Out_Identity 2 2 Input_2 -0.090580426 0.1858058 UNIT_INPUT 2 0 0 Act_Identity Out_Identity 3 3 Input_3 0.823582947 -0.2467227 UNIT_INPUT 3 0 0 Act_Identity Out_Identity 4 4 Input_4 1.077348471 -0.2271125 UNIT_INPUT 4 0 0 Act_Identity Out_Identity 5 5 Hidden_2_1 0.827119827 -3.0416200 UNIT_HIDDEN 1 2 0 Act_Logistic Out_Identity 6 6 Hidden_2_2 0.979820251 0.8963667 UNIT_HIDDEN 2 2 0 Act_Logistic Out_Identity 7 7 Hidden_2_3 0.004135765 -1.5327449 UNIT_HIDDEN 3 2 0 Act_Logistic Out_Identity 8 8 Hidden_2_4 0.018411839 -0.9707544 UNIT_HIDDEN 4 2 0 Act_Logistic Out_Identity 9 9 Hidden_2_5 0.171753630 2.6582935 UNIT_HIDDEN 5 2 0 Act_Logistic Out_Identity 10 10 Output_1 0.013317095 -1.1436981 UNIT_OUTPUT 1 4 0 Act_Logistic Out_Identity 11 11 Output_2 0.129099220 -0.6441937 UNIT_OUTPUT 2 4 0 Act_Logistic Out_Identity 12 12 Output_3 0.881433606 -1.2297773 UNIT_OUTPUT 3 4 0 Act_Logistic Out_Identity $fullWeightMatrix Input_1 Input_2 Input_3 Input_4 Hidden_2_1 Hidden_2_2 Hidden_2_3 Hidden_2_4 Hidden_2_5 Output_1 Output_2 Output_3 Input_1 0 0 0 0 -0.3508658 0.2684588 -0.5913861 -0.5612198 0.3195967 0.00000000 0.000000 0.000000 Input_2 0 0 0 0 -0.5082296 -0.9437067 1.2823328 0.9810203 0.2950290 0.00000000 0.000000 0.000000 Input_3 0 0 0 0 2.2801065 1.2012331 -1.8615422 -1.2803886 -1.9130263 0.00000000 0.000000 0.000000 Input_4 0 0 0 0 2.7568083 1.5704848 -1.6876845 -1.3023156 -2.6831374 0.00000000 0.000000 0.000000 Hidden_2_1 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -1.40362704 -3.796962 3.113594 Hidden_2_2 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -2.07435131 1.398071 1.325584 Hidden_2_3 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 2.11041474 -3.148410 -2.083688 Hidden_2_4 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 1.70966542 -1.791131 -2.028595 Hidden_2_5 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 -0.04861032 3.213499 -3.448702 Output_1 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 Output_2 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 Output_3 0 0 0 0 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.00000000 0.000000 0.000000 *laser library(RSNNS) data(snnsData) inputs <- snnsData$laser_1000.pat[,inputColumns(snnsData$laser_1000.pat)] outputs <- snnsData$laser_1000.pat[,outputColumns(snnsData$laser_1000.pat)] patterns <- splitForTrainingAndTest(inputs, outputs, ratio=0.15) model <- elman(patterns$inputsTrain, patterns$targetsTrain, size=c(8,8), learnFuncParams=c(0.1), maxit=500, inputsTest=patterns$inputsTest, targetsTest=patterns$targetsTest, linOut=FALSE) modelJordan <- jordan(patterns$inputsTrain, patterns$targetsTrain, size=c(8), learnFuncParams=c(0.1), maxit=100, inputsTest=patterns$inputsTest, targetsTest=patterns$targetsTest, linOut=FALSE) modelMlp <- mlp(patterns$inputsTrain, patterns$targetsTrain, initFuncParams=c(-0.3,0.3),size=c(8), learnFuncParams=c(0.05), maxit=500, inputsTest=patterns$inputsTest, targetsTest=patterns$targetsTest, linOut=TRUE) names(model) par(mfrow=c(3,3)) plotIterativeError(model) plotIterativeError(modelJordan) #plotIterativeError(modelMlp) plotRegressionError(patterns$targetsTrain, model$fitted.values, main="Regression Plot Fit") plotRegressionError(patterns$targetsTest, model$fittedTestValues, main="Regression Plot Test") hist(model$fitted.values - patterns$targetsTrain, col="lightblue", main="Error Histogram Fit") #model$IterativeFitError[length(model$IterativeFitError)] plot(inputs, type="l") plot(inputs[1:100], type="l") lines(outputs[1:100], col="red") lines(model$fitted.values[1:100], col="green") *som library(RSNNS) #data(snnsData) #inputs <- snnsData$som_cube.pat data(iris) inputs <- normalizeData(iris[,1:4], "norm") model <- som(inputs, mapX=16, mapY=16, maxit=500, calculateActMaps=TRUE, targets=iris[,5]) par(mfrow=c(3,3)) for(i in 1:ncol(inputs)) plotActMap(model$componentMaps[[i]], col=rev(topo.colors(12))) plotActMap(model$map, col=rev(heat.colors(12))) plotActMap(log(model$map+1), col=rev(heat.colors(12))) persp(1:model$archParams$mapX, 1:model$archParams$mapY, log(model$map+1), theta = 30, phi = 30, expand = 0.5, col = "lightblue") plotActMap(model$labeledMap) # 各セルがどのラベルを持つか model$componentMaps # 各ファクター（irisなら４ファクター）がマップ全体（各セル）にどのように分布しているかを示す model$labeledUnits # 各セルにどのクラスが存在しているか．model$mapの内訳 model$map # 各セルに存在するデータの個数 names(model) *assoz dataset <- snnsData$dlvq_ziff_100.pat patterns <- dataset[,inputColumns(dataset)][c(1,2,3,4,5,6,8,10,11,12),] model <- assoz(patterns, dimX=16, dimY=16) #model$fitted.values par(mfrow=c(3,3)) for (i in 1:9) plotActMap(matrixToActMapList(patterns,16)[[i]]) actMaps <- matrixToActMapList(model$fitted.values, nrow=16) for (i in 1:9) plotActMap(actMaps[[i]]) pred <- predict(model, dataset[,inputColumns(dataset)]) for (i in 1:length(pred)) plotActMap(pred[[i]]) *assoz ## Not run: demo(assoz_letters) ## Not run: demo(assoz_lettersSnnsR) data(snnsData) patterns <- snnsData$art1_letters.pat model <- assoz(patterns, dimX=7, dimY=5) actMaps <- matrixToActMapList(model$fitted.values, nrow=7) par(mfrow=c(3,3)) for (i in 1:9) plotActMap(actMaps[[i]]) *assoz demo(assoz_letters) library(RSNNS) data(snnsData) patterns <- snnsData$art1_letters.pat model <- assoz(patterns, dimX=7, dimY=5) #model$fitted.values par(mfrow=c(3,3)) for (i in 1:9) plotActMap(model$fitted.values[[i]]) cbind(patterns[1,], as.vector(t(model$fitted[[1]])), predict(model, patterns)[1,]) testData <- rbind( as.vector(t(rbind(c(0,0,0,0,1), c(0,0,0,1,0), c(0,0,0,1,0), c(0,0,1,0,0), c(0,1,0,0,0), c(0,1,0,0,0), c(1,0,0,0,0)))), as.vector(t(rbind(c(1,0,0,0,0), c(1,0,0,0,0), c(0,1,0,0,0), c(0,0,1,0,0), c(0,0,0,1,0), c(0,0,0,0,1), c(0,0,0,0,1)))) ) model <- assoz(patterns[1:5,], dimX=7, dimY=5, maxit=1000) testData <- rbind(patterns[1,], patterns[1,]) testData[2,1] <- 1 testData[2,2] <- 0 cbind(patterns[1,], as.vector(t(model$fitted[[1]])), testData[2,], predict(model, testData)[2,]) par(mfrow=c(2,2)) pred <- predict(model, testData) plotActMap(matrixToActMapList(patterns,7)[[1]]) plotActMap(model$fitted[[1]]) plotActMap(matrixToActMapList(testData,7)[[2]]) plotActMap(matrixToActMapList(pred,7)[[2]]) pred <- predict(model, pred) plotActMap(matrixToActMapList(testData,7)[[1]]) plotActMap(matrixToActMapList(pred,7)[[1]]) plotActMap(matrixToActMapList(testData,7)[[2]]) plotActMap(matrixToActMapList(pred,7)[[2]]) !!コメント //{{comment}}