Export - CSV (All fields)
Export - CSV (Current fields)
SPARK-6337
Spark 1.3 doc fixes
SPARK-6336
LBFGS should document what convergenceTol means- SPARK-6455
Correct some mistakes and typos - SPARK-6657
Fix Python doc build warnings SPARK-6597
Replace `input:checkbox` with `input[type="checkbox"] in additional-metrics.js
ML doc cleanup: logreg, classification link
Add support for Intersect and Except in Catalyst DSL- SPARK-7723
Fix string interpolation in pipeline examples
Fixing broken trainImplicit example in MLlib Collaborative Filtering documentation.
Use reservoir sample in RandomForest when choosing features per node
Update plugin versions in Maven build and centralize versions
HashingTF should output ML attributes
Add Matrix, SparseMatrix to __all__ list in linalg.py
DecisionTree: eliminate bins for unordered features
Scala NaiveBayes should expose getLambda
Mark parts of LBFGS, GradientDescent as DeveloperApi
Rename `Math` to `math` in MLlib's Scala code- SPARK-7427
Make sharedParams match in Scala, Python
Provide an easier way for developers to handle deprecated configs
Make mesosExecutorCores configurable in mesos "fine-grained" mode
Support jdbc connection properties in OPTIONS part of the query
DriverRunner doesn't support pulling HTTP/HTTPS URIsSPARK-5727
Deprecate, remove Debian packaging
Optimize LogisticRegressionModel.predictPoint
Deprecate jsonRDD and replace it by jsonDataFrame / jsonDF
Use while(true) in OpenHashSet.getPos
Add contains to OpenHashMap
Remove the reflection call in HiveFunctionWrapper- SPARK-6469
Improving documentation on YARN local directories usage
Python type errors should print type, not object
Python KMeans paritySPARK-6608
Make DataFrame.rdd a lazy val
Make DataFrame.where accept a string conditionExpr
UIWorkloadGenerator: The main thread should not stop SparkContext until all jobs finishSPARK-4804
StringContext method to allow using Strings for column names in catalyst DSL
Categorical feature indexing
ElementwiseProduct scalingVec param name should match between ml,mllib
Scheduler delay shown in event timeline is incorrect
Ignores files whose name starts with "." while enumerating files in HadoopFsRelation
Improve Evaluator doc to state that higher metric values are better.
Add sqlContext to MLlibTestSparkContext
NaiveBayes.modelType should use lowercase letters
VectorIndexer: do not preserve non-ML metadata in output- SPARK-7255
spark.streaming.kafka.maxRetries not documented
typo spark.executor.extraLibraryOptions => spark.executor.extraLibraryPath
Use getOrElse for getting the threshold of LR model
SPARK building documentation still mentions building for yarn 0.23
Use itertools.product in ParamGridBuilder
Wrap FPGrowthModel.freqItemsets with namedtuples (or document the return type) in PySpark
Bucketizer should include the right most boundary in the last bucket.