flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-1933) Add distance measure interface and basic implementation to machine learning library
Date Thu, 07 May 2015 09:00:10 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14532283#comment-14532283
] 

ASF GitHub Bot commented on FLINK-1933:
---------------------------------------

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

    https://github.com/apache/flink/pull/629#discussion_r29834901
  
    --- Diff: flink-staging/flink-ml/src/main/scala/org/apache/flink/ml/math/metrics/distances/CosineDistanceMeasure.scala
---
    @@ -0,0 +1,45 @@
    +/*
    + * 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.flink.ml.math.metrics.distances
    +
    +import org.apache.flink.ml.math.Vector
    +
    +/** This class implements a cosine distance metric. The class calculates the distance
between
    +  * the given vectors by dividing the dot product of two vectors by the product of their
lengths.
    +  * We convert the result of division to a usable distance. So, 1 - cos(angle) is actually
returned.
    +  *
    +  * @see http://en.wikipedia.org/wiki/Cosine_similarity
    +  */
    +class CosineDistanceMeasure extends DistanceMeasure {
    +  override def distance(a: Vector, b: Vector): Double = {
    +    checkValidArguments(a, b)
    +
    +    val dotProd: Double = a.dot(b)
    +    val denominator: Double = a.magnitude * b.magnitude
    +    if (dotProd == 0 && denominator == 0) {
    --- End diff --
    
    what if `a` and `b` are both zero? Are they then similar with respect to the cosine similarity?


> Add distance measure interface and basic implementation to machine learning library
> -----------------------------------------------------------------------------------
>
>                 Key: FLINK-1933
>                 URL: https://issues.apache.org/jira/browse/FLINK-1933
>             Project: Flink
>          Issue Type: New Feature
>          Components: Machine Learning Library
>            Reporter: Chiwan Park
>            Assignee: Chiwan Park
>              Labels: ML
>
> Add distance measure interface to calculate distance between two vectors and some implementations
of the interface. In FLINK-1745, [~till.rohrmann] suggests a interface following:
> {code}
> trait DistanceMeasure {
>   def distance(a: Vector, b: Vector): Double
> }
> {code}
> I think that following list of implementation is sufficient to provide first to ML library
users.
> * Manhattan distance [1]
> * Cosine distance [2]
> * Euclidean distance (and Squared) [3]
> * Tanimoto distance [4]
> * Minkowski distance [5]
> * Chebyshev distance [6]
> [1]: http://en.wikipedia.org/wiki/Taxicab_geometry
> [2]: http://en.wikipedia.org/wiki/Cosine_similarity
> [3]: http://en.wikipedia.org/wiki/Euclidean_distance
> [4]: http://en.wikipedia.org/wiki/Jaccard_index#Tanimoto_coefficient_.28extended_Jaccard_coefficient.29
> [5]: http://en.wikipedia.org/wiki/Minkowski_distance
> [6]: http://en.wikipedia.org/wiki/Chebyshev_distance



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message