[GitHub] spark pull request: [SPARK-2511][MLLIB] add HashingTF and IDF

[MLnick <gi...@git.apache.org>]

Github user MLnick commented on a diff in the pull request:

    https://github.com/apache/spark/pull/1671#discussion_r15775472
  
    --- Diff: mllib/src/main/scala/org/apache/spark/mllib/feature/HashingTF.scala ---
    @@ -0,0 +1,79 @@
    +/*
    + * 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.feature
    +
    +import java.lang.{Iterable => JavaIterable}
    +
    +import scala.collection.JavaConverters._
    +import scala.collection.mutable
    +
    +import org.apache.spark.annotation.Experimental
    +import org.apache.spark.api.java.JavaRDD
    +import org.apache.spark.mllib.linalg.{Vector, Vectors}
    +import org.apache.spark.rdd.RDD
    +import org.apache.spark.util.Utils
    +
    +/**
    + * :: Experimental ::
    + * Maps a sequence of terms to their term frequencies using the hashing trick.
    + *
    + * @param numFeatures number of features (default: 1000000)
    + */
    +@Experimental
    +class HashingTF(val numFeatures: Int) extends Serializable {
    +
    +  def this() = this(1000000)
    +
    --- End diff --
    
    @mengxr @mateiz this looks really awesome!
    
    I know this is merged already, but one comment: when using the hashing trick, should the vector size not usually be a power of 2? This is mentioned [here](http://scikit-learn.org/stable/modules/feature_extraction.html#implementation-details) for example.
    
    Pretty much every library uses powers of 2 - Vowpal Wabbit, sophia-ml, scikit-learn and shogun for example.
    
    So it may be worth mentioning, and making the default 2^20 (or 2^18 which is also a common default).


---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org