From 1537e55604cafafa49a8b7f3ce915f9745392bc0 Mon Sep 17 00:00:00 2001 From: Kai Jiang Date: Tue, 5 Jan 2016 15:33:27 -0800 Subject: [SPARK-12041][ML][PYSPARK] Add columnSimilarities to IndexedRowMatrix Add `columnSimilarities` to IndexedRowMatrix for PySpark spark.mllib.linalg. Author: Kai Jiang Closes #10158 from vectorijk/spark-12041. --- python/pyspark/mllib/linalg/distributed.py | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'python/pyspark') diff --git a/python/pyspark/mllib/linalg/distributed.py b/python/pyspark/mllib/linalg/distributed.py index 0e76050788..e1f022187d 100644 --- a/python/pyspark/mllib/linalg/distributed.py +++ b/python/pyspark/mllib/linalg/distributed.py @@ -297,6 +297,20 @@ class IndexedRowMatrix(DistributedMatrix): """ return self._java_matrix_wrapper.call("numCols") + def columnSimilarities(self): + """ + Compute all cosine similarities between columns. + + >>> rows = sc.parallelize([IndexedRow(0, [1, 2, 3]), + ... IndexedRow(6, [4, 5, 6])]) + >>> mat = IndexedRowMatrix(rows) + >>> cs = mat.columnSimilarities() + >>> print(cs.numCols()) + 3 + """ + java_coordinate_matrix = self._java_matrix_wrapper.call("columnSimilarities") + return CoordinateMatrix(java_coordinate_matrix) + def toRowMatrix(self): """ Convert this matrix to a RowMatrix. -- cgit v1.2.3