XGboost is now an option in pyspark pipelines (link here) but PysparkML also supports a number of alternatives including

1. Gradient-Boosted Trees (GBTs): Gradient-boosted trees is another boosting ensemble technique that learns from its mistakes in previous iterations. For this one can use the GBTClassifier from pyspark.ml.classification.
2. Random Forest: Random Forest is a bagging ensemble learning method. For this one can use the RandomForestClassifier from pyspark.ml.classification.
3. Decision Tree Classifier: The decision tree is a simple yet effective machine learning algorithm. For this one can use theDecisionTreeClassifier from pyspark.ml.classification.