Spark Collaborative Filtering Learner

This node utilizes the Apache Spark collaborative filtering implementation.

Notice: The matrix factorization model contains references to Spark DataFrame/RDDs and thus is not selfcontained. The referenced Spark DataFrame/RDDs will be deleted, like any other Spark DataFrame/RDD generated in KNIME, when the node is reset, or when the workflow is closed and the "Delete Spark DataFrame/RDDs on dispose" option is enabled for the current Spark context.

Options

User column
The user column. Supports long and integer columns, however the node will fail, if the actual values are outside of the integer range, due to limitations of Spark MLlib.
Product column
The product column. Supports long and integer columns, however the node will fail, if the actual values are outside of the integer range, due to limitations of Spark MLlib.
Rating column
The rating column that contains the rating of the user for the given product. Supports numeric columns.
Lambda
Specifies the regularization parameter in ALS.
Alpha
Is a parameter applicable to the implicit feedback variant of ALS that governs the baseline confidence in preference observations.
Rank
The number of latent factors in the model.
Iterations
The number of iterations to run.
Number of blocks
The number of blocks used to parallelize computation (set to -1 to auto-configure).
Implicit feedback
Select this option to use the ALS variant adapted for implicit feedback data otherwise the explicit feedback ALS variant is used.
Initialization seed
Random seed to initialize the factors.

Input Ports

Icon
Spark DataFrame/RDD with user ratings of products

Output Ports

Icon
The input data labeled with the prediction.
Icon
The matrix factorization model. The model is not selfcontained and thus can not be saved to file.

Views

This node has no views

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.