diff options
author | Sean Owen <sowen@cloudera.com> | 2016-10-03 18:09:36 +0000 |
---|---|---|
committer | DB Tsai <dbtsai@dbtsai.com> | 2016-10-03 18:09:36 +0000 |
commit | 1dd68d3827133d203e85294405400b04904879e0 (patch) | |
tree | 0e4e81383789242552cc93a40a95c05d5322e122 | |
parent | 7bf92127643570e4eb3610fa3ffd36839eba2718 (diff) | |
download | spark-1dd68d3827133d203e85294405400b04904879e0.tar.gz spark-1dd68d3827133d203e85294405400b04904879e0.tar.bz2 spark-1dd68d3827133d203e85294405400b04904879e0.zip |
[SPARK-17718][DOCS][MLLIB] Make loss function formulation label note clearer in MLlib docs
## What changes were proposed in this pull request?
Move note about labels being +1/-1 in formulation only to be just under the table of formulations.
## How was this patch tested?
Doc build
Author: Sean Owen <sowen@cloudera.com>
Closes #15330 from srowen/SPARK-17718.
-rw-r--r-- | docs/mllib-linear-methods.md | 9 |
1 files changed, 5 insertions, 4 deletions
diff --git a/docs/mllib-linear-methods.md b/docs/mllib-linear-methods.md index 6fcd3ae857..816bdf1317 100644 --- a/docs/mllib-linear-methods.md +++ b/docs/mllib-linear-methods.md @@ -78,6 +78,11 @@ methods `spark.mllib` supports: </tbody> </table> +Note that, in the mathematical formulation above, a binary label $y$ is denoted as either +$+1$ (positive) or $-1$ (negative), which is convenient for the formulation. +*However*, the negative label is represented by $0$ in `spark.mllib` instead of $-1$, to be consistent with +multiclass labeling. + ### Regularizers The purpose of the @@ -136,10 +141,6 @@ multiclass classification problems. For both methods, `spark.mllib` supports L1 and L2 regularized variants. The training data set is represented by an RDD of [LabeledPoint](mllib-data-types.html) in MLlib, where labels are class indices starting from zero: $0, 1, 2, \ldots$. -Note that, in the mathematical formulation in this guide, a binary label $y$ is denoted as either -$+1$ (positive) or $-1$ (negative), which is convenient for the formulation. -*However*, the negative label is represented by $0$ in `spark.mllib` instead of $-1$, to be consistent with -multiclass labeling. ### Linear Support Vector Machines (SVMs) |