Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/964#discussion_r13742971
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/linalg/EigenValueDecomposition.scala
---
@@ -0,0 +1,150 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.mllib.linalg
+
+import breeze.linalg.{DenseMatrix => BDM, DenseVector => BDV}
+import com.github.fommil.netlib.ARPACK
+import org.netlib.util.{intW, doubleW}
+
+import org.apache.spark.annotation.Experimental
+
+/**
+ * :: Experimental ::
+ * Represents eigenvalue decomposition factors.
+ */
+@Experimental
+case class EigenValueDecomposition[VType](s: Vector, V: VType)
+
+@Experimental
+object EigenValueDecomposition {
+ /**
+ * Compute the leading k eigenvalues and eigenvectors on a symmetric
square matrix using ARPACK.
+ * The caller needs to ensure that the input matrix is real symmetric.
This function requires
+ * memory for `n*(4*k+4)` doubles.
+ *
+ * @param mul a function that multiplies the symmetric matrix with a
DenseVector.
+ * @param n dimension of the square matrix (maximum Int.MaxValue).
+ * @param k number of leading eigenvalues required.
+ * @param tol tolerance of the eigs computation.
+ * @return a dense vector of eigenvalues in descending order and a dense
matrix of eigenvectors
+ * (columns of the matrix). The number of computed eigenvalues
might be smaller than k.
+ */
+ private[mllib] def symmetricEigs(mul: DenseVector => DenseVector, n:
Int, k: Int, tol: Double)
+ : (BDV[Double], BDM[Double]) = {
+ // TODO: remove this function and use eigs in breeze when switching
breeze version
+ require(n > k, s"Number of required eigenvalues $k must be smaller
than matrix dimension $n")
+
+ val arpack = ARPACK.getInstance()
+
+ // tolerance used in stopping criterion
+ val tolW = new doubleW(tol)
+ // number of desired eigenvalues, 0 < nev < n
+ val nev = new intW(k)
+ // nev Lanczos vectors are generated are generated in the first
iteration
--- End diff --
remove "are generated"
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
