From 789a0f5b8b441e5d3aa4d2f475b39ab5cabcb112 Mon Sep 17 00:00:00 2001
From: Bago Amirbekian <bago@databricks.com>
Date: Wed, 30 May 2018 08:18:49 -0700
Subject: [PATCH] Added benchmarks to mllib-large.yaml for classifcation
 Estimators. (#143)

---
 .../sql/perf/mllib/config/mllib-large.yaml    | 26 +++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/src/main/resources/com/databricks/spark/sql/perf/mllib/config/mllib-large.yaml b/src/main/resources/com/databricks/spark/sql/perf/mllib/config/mllib-large.yaml
index ebcc24b..15d4528 100644
--- a/src/main/resources/com/databricks/spark/sql/perf/mllib/config/mllib-large.yaml
+++ b/src/main/resources/com/databricks/spark/sql/perf/mllib/config/mllib-large.yaml
@@ -11,3 +11,29 @@ benchmarks:
     params:
       depth: [5, 10]
       numClasses: 4
+  - name: classification.GBTClassification
+    params:
+      numFeatures: 3000
+      depth: 5
+      numClasses: 4
+      maxIter: 10
+  - name: classification.RandomForestClassification
+    params:
+      depth: 10
+      numClasses: 4
+      maxIter: 200 # number of trees
+  - name: classification.LogisticRegression
+    params:
+      regParam: 0.01
+      tol: 0.0
+      maxIter: 20
+  - name: classification.LinearSVC
+    params:
+      regParam: 0.01
+      tol: 0
+      maxIter: 20
+  - name: classification.NaiveBayes
+    params:
+      numFeatures: 5000
+      numClasses: 2
+      smoothing: 1.0