public class CollaborativeFiltering extends Object implements Serializable
Constructor and Description |
---|
CollaborativeFiltering() |
Modifier and Type | Method and Description |
---|---|
static List<Integer> |
recommendProducts(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model,
int userId,
int numberOfProducts)
This method recommends products for a given user.
|
static List<Integer> |
recommendUsers(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model,
int productId,
int numberOfUsers)
This method recommends users for a given product.
|
org.apache.spark.api.java.JavaDoubleRDD |
test(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model,
org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> testData) |
org.apache.spark.mllib.recommendation.MatrixFactorizationModel |
trainExplicit(org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> trainingDataset,
int rank,
int noOfIterations,
double regularizationParameter,
int noOfBlocks)
This method uses alternating least squares (ALS) algorithm to train a matrix factorization model given an JavaRDD
of ratings given by users to some products.
|
org.apache.spark.mllib.recommendation.MatrixFactorizationModel |
trainImplicit(org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> trainingDataset,
int rank,
int noOfIterations,
double regularizationParameter,
double confidenceParameter,
int noOfBlocks)
This method uses alternating least squares (ALS) algorithm to train a matrix factorization model given an JavaRDD
of 'implicit preferences' given by users to some products.
|
public org.apache.spark.mllib.recommendation.MatrixFactorizationModel trainExplicit(org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> trainingDataset, int rank, int noOfIterations, double regularizationParameter, int noOfBlocks)
trainingDataset
- Training dataset as a JavaRDD of Ratingsrank
- Number of latent factorsnoOfIterations
- Number of iterationsregularizationParameter
- Regularization parameternoOfBlocks
- Level of parallelism (auto configure = -1)public org.apache.spark.mllib.recommendation.MatrixFactorizationModel trainImplicit(org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> trainingDataset, int rank, int noOfIterations, double regularizationParameter, double confidenceParameter, int noOfBlocks)
trainingDataset
- Training dataset as a JavaRDD of Ratingsrank
- Number of latent factorsnoOfIterations
- Number of iterationsregularizationParameter
- Regularization parameterconfidenceParameter
- Confidence parameternoOfBlocks
- Level of parallelism (auto configure = -1)public org.apache.spark.api.java.JavaDoubleRDD test(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model, org.apache.spark.api.java.JavaRDD<org.apache.spark.mllib.recommendation.Rating> testData)
public static List<Integer> recommendProducts(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model, int userId, int numberOfProducts) throws MLModelHandlerException
model
- Matrix factorization modeluserId
- The user to recommend products tonumberOfProducts
- Number of products to returnMLModelHandlerException
public static List<Integer> recommendUsers(org.apache.spark.mllib.recommendation.MatrixFactorizationModel model, int productId, int numberOfUsers) throws MLModelHandlerException
model
- Matrix factorization modelproductId
- The product to recommend users tonumberOfUsers
- Number of users to returnMLModelHandlerException
Copyright © 2016 WSO2, Inc.. All Rights Reserved.