scripts.algorithms.ALS_predict.dml Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of systemml Show documentation
Show all versions of systemml Show documentation
Declarative Machine Learning
#-------------------------------------------------------------
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
#
#-------------------------------------------------------------
#
# THIS SCRIPT COMPUTES THE RATING/SCORE FOR A GIVEN LIST OF PAIRS: (USER-ID, ITEM-ID) USING 2 FACTOR MATRICES L AND R
# WE ASSUME THAT ALL USERS HAVE RATED AT LEAST ONCE AND ALL ITEMS HAVE BEEN RATED AT LEAST ONCE.
# INPUT PARAMETERS:
# ---------------------------------------------------------------------------------------------
# NAME TYPE DEFAULT MEANING
# ---------------------------------------------------------------------------------------------
# X String --- The input user-id/item-id list
# Y String --- The output user-id/item-id/score
# L String --- Location of the factor matrix L: user-id x feature-id
# R String --- Location to the factor matrix R: feature-id x item-id
# Vrows Integer --- The number of rows in the original matrix
# Vcols Integer --- The number of columns in the original matrix
# fmt String "text" The output format of the factor matrix user-id/item-id/score
# ---------------------------------------------------------------------------------------------
# OUTPUT: Matrix Y containing the predicted ratings for users and items specified in input matrix X
#
# HOW TO INVOKE THIS SCRIPT - EXAMPLE:
# hadoop jar SystemML.jar -f ALS-predict.dml -nvargs Vrows=100000 Vcols=10000 X=INPUT_DIR/X L=OUTPUT_DIR/L R=OUTPUT_DIR/R
# Y=OUTPUT_DIR/Y fmt=csv
fileX = $X;
fileY = $Y;
fileL = $L;
fileR = $R;
Vrows = $Vrows;
Vcols = $Vcols;
fmtO = ifdef ($fmt, "text"); # $fmt="text";
X = read (fileX);
L = read (fileL);
R = read (fileR);
# we assume that the number of columns in X is equal to 2: user-id and item-id
n = nrow (X);
m = ncol (X);
if (m != 2){
stop("The input matrix must have 2 columns: user-id and item-id");
}
Lrows = nrow (L);
Rcols = ncol (R);
X_user_max = max (X[,1]);
X_item_max = max (X[,2]);
# initializing Y matrix
Y = matrix(0, rows = n, cols = 3);
if (X_user_max > Vrows | X_item_max > Vcols ) {
stop ("Predictions cannot be provided. Maximum user-id (item-id) exceed the number of rows (columns) of V.");
}
if (Lrows != Vrows | Rcols != Vcols) {
stop ("Predictions cannot be provided. Number of rows of L (columns of R) does not match the number of rows (column) of V.");
}
# user2item table
ones = matrix (1, rows = n, cols = 1);
UI = table (X[,1], X[,2], ones, Vrows, Vcols);
# summing up over all items for all users
U = rowSums (UI)
# replacing all rows > 1 with 1
U = ppred (U, 1, ">=");
# selecting users from factor L
U_prime = L * U;
V_prime = (U_prime %*% R);
# Applying items filter
V_prime = UI * V_prime;
write(V_prime, fileY, format = fmtO);