[SPARK-18007][SPARKR][ML] update SparkR MLP - add initalWeights parameter

## What changes were proposed in this pull request?

update SparkR MLP, add initalWeights parameter.

## How was this patch tested?

test added.

Author: WeichenXu <WeichenXu123@outlook.com>

Closes #15552 from WeichenXu123/mlp_r_add_initialWeight_param.

R/pkg/R/mllib.R

@@ -665,6 +665,8 @@ setMethod("predict", signature(object = "KMeansModel"),
 #' @param tol convergence tolerance of iterations.
 #' @param stepSize stepSize parameter.
 #' @param seed seed parameter for weights initialization.
+#' @param initialWeights initialWeights parameter for weights initialization, it should be a
+#' numeric vector.
 #' @param ... additional arguments passed to the method.
 #' @return \code{spark.mlp} returns a fitted Multilayer Perceptron Classification Model.
 #' @rdname spark.mlp
@@ -677,8 +679,9 @@ setMethod("predict", signature(object = "KMeansModel"),
 #' df <- read.df("data/mllib/sample_multiclass_classification_data.txt", source = "libsvm")
 #'
 #' # fit a Multilayer Perceptron Classification Model
-#' model <- spark.mlp(df, blockSize = 128, layers = c(4, 5, 4, 3), solver = "l-bfgs",
-#'                    maxIter = 100, tol = 0.5, stepSize = 1, seed = 1)
+#' model <- spark.mlp(df, blockSize = 128, layers = c(4, 3), solver = "l-bfgs",
+#'                    maxIter = 100, tol = 0.5, stepSize = 1, seed = 1,
+#'                    initialWeights = c(0, 0, 0, 0, 0, 5, 5, 5, 5, 5, 9, 9, 9, 9, 9))
 #'
 #' # get the summary of the model
 #' summary(model)
@@ -695,7 +698,7 @@ setMethod("predict", signature(object = "KMeansModel"),
 #' @note spark.mlp since 2.1.0
 setMethod("spark.mlp", signature(data = "SparkDataFrame"),
           function(data, layers, blockSize = 128, solver = "l-bfgs", maxIter = 100,
-                   tol = 1E-6, stepSize = 0.03, seed = NULL) {
+                   tol = 1E-6, stepSize = 0.03, seed = NULL, initialWeights = NULL) {
             if (is.null(layers)) {
               stop ("layers must be a integer vector with length > 1.")
             }
@@ -706,10 +709,13 @@ setMethod("spark.mlp", signature(data = "SparkDataFrame"),
             if (!is.null(seed)) {
               seed <- as.character(as.integer(seed))
             }
+            if (!is.null(initialWeights)) {
+              initialWeights <- as.array(as.numeric(na.omit(initialWeights)))
+            }
             jobj <- callJStatic("org.apache.spark.ml.r.MultilayerPerceptronClassifierWrapper",
                                 "fit", data@sdf, as.integer(blockSize), as.array(layers),
                                 as.character(solver), as.integer(maxIter), as.numeric(tol),
-                                as.numeric(stepSize), seed)
+                                as.numeric(stepSize), seed, initialWeights)
             new("MultilayerPerceptronClassificationModel", jobj = jobj)
           })
 

R/pkg/inst/tests/testthat/test_mllib.R

@@ -410,6 +410,21 @@ test_that("spark.mlp", {
   model <- spark.mlp(df, layers = c(4, 5, 4, 3), maxIter = 10, seed = 10)
   mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
   expect_equal(head(mlpPredictions$prediction, 12), c(1, 1, 1, 1, 2, 1, 2, 2, 1, 0, 0, 1))
+
+  # test initialWeights
+  model <- spark.mlp(df, layers = c(4, 3), maxIter = 2, initialWeights =
+    c(0, 0, 0, 0, 0, 5, 5, 5, 5, 5, 9, 9, 9, 9, 9))
+  mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
+  expect_equal(head(mlpPredictions$prediction, 12), c(1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1))
+
+  model <- spark.mlp(df, layers = c(4, 3), maxIter = 2, initialWeights =
+    c(0.0, 0.0, 0.0, 0.0, 0.0, 5.0, 5.0, 5.0, 5.0, 5.0, 9.0, 9.0, 9.0, 9.0, 9.0))
+  mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
+  expect_equal(head(mlpPredictions$prediction, 12), c(1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1))
+
+  model <- spark.mlp(df, layers = c(4, 3), maxIter = 2)
+  mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
+  expect_equal(head(mlpPredictions$prediction, 12), c(1, 1, 1, 1, 0, 1, 0, 2, 1, 0, 0, 1))
 })
 
 test_that("spark.naiveBayes", {

mllib/src/main/scala/org/apache/spark/ml/r/MultilayerPerceptronClassifierWrapper.scala

@@ -24,6 +24,7 @@ import org.json4s.jackson.JsonMethods._
 
 import org.apache.spark.ml.{Pipeline, PipelineModel}
 import org.apache.spark.ml.classification.{MultilayerPerceptronClassificationModel, MultilayerPerceptronClassifier}
+import org.apache.spark.ml.linalg.Vectors
 import org.apache.spark.ml.util.{MLReadable, MLReader, MLWritable, MLWriter}
 import org.apache.spark.sql.{DataFrame, Dataset}
 
@@ -58,7 +59,8 @@ private[r] object MultilayerPerceptronClassifierWrapper
       maxIter: Int,
       tol: Double,
       stepSize: Double,
-      seed: String
+      seed: String,
+      initialWeights: Array[Double]
      ): MultilayerPerceptronClassifierWrapper = {
     // get labels and feature names from output schema
     val schema = data.schema
@@ -73,6 +75,11 @@ private[r] object MultilayerPerceptronClassifierWrapper
       .setStepSize(stepSize)
       .setPredictionCol(PREDICTED_LABEL_COL)
     if (seed != null && seed.length > 0) mlp.setSeed(seed.toInt)
+    if (initialWeights != null) {
+      require(initialWeights.length > 0)
+      mlp.setInitialWeights(Vectors.dense(initialWeights))
+    }
+
     val pipeline = new Pipeline()
       .setStages(Array(mlp))
       .fit(data)