partial function for prediction, auxiliary function h2o github
partial function for prediction, auxiliary function h2o github
Predict on Row, one Row is one instance, helpinng function, used in predictRow
Predict on Row, one Row is one instance, helpinng function, used in predictRow
one instance
Row column names (dataframe column names)
label-score pair (if applicable)
feature and response column names
feature and response column names
predict on each RowData
predict on each RowData
Tuple2[String, Double]: String: class/cluster/low-dimension-string/regression-value-string Double: confidence score if applicable
prediction on a whole dataframe with key attached
prediction on a whole dataframe with key attached
data
keys, default=Array()
the column name used to store the prediction label, default="label"
the column name used to store the prediction score, default="score"
dataframe with key(s), label, score
Predict on RDD
Predict on RDD
One instance is a RDD[Tuple2[String, Any]], where each element is a feature (name-value pair)
label-score pairs (if applicable)
Prediction on Row data with key attached
Prediction on Row data with key attached
data
data column names
keys, array of strings, kept in the result, default=Array()
a row of predicted label with confidence score (if applicalbe) and original keys (if provided)
Conditional Prediction on Row data with key attached, along with non-prediction value/score
Conditional Prediction on Row data with key attached, along with non-prediction value/score
data
data column names
the logic function to decide whether apply the model on this particular instance
origin label column, if not applying the model, this column's value will be directy retrieved
origin keys, default=Array()
the row of predictions with keys (optional)
H2OSparkPredictor
A trained H2O Model for use in Spark Environment. It loads a pre-trained h2o pojo/mojo model and add prediction functions on RDD, Row, DataFrame