1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
|
/*
* 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.
*/
package org.apache.spark.graphx.lib
import scala.util.Random
import com.github.fommil.netlib.BLAS.{getInstance => blas}
import org.apache.spark.graphx._
import org.apache.spark.rdd._
/** Implementation of SVD++ algorithm. */
object SVDPlusPlus {
/** Configuration parameters for SVDPlusPlus. */
class Conf(
var rank: Int,
var maxIters: Int,
var minVal: Double,
var maxVal: Double,
var gamma1: Double,
var gamma2: Double,
var gamma6: Double,
var gamma7: Double)
extends Serializable
/**
* This method is now replaced by the updated version of `run()` and returns exactly
* the same result.
*/
@deprecated("Call run()", "1.4.0")
def runSVDPlusPlus(edges: RDD[Edge[Double]], conf: Conf)
: (Graph[(Array[Double], Array[Double], Double, Double), Double], Double) =
{
run(edges, conf)
}
/**
* Implement SVD++ based on "Factorization Meets the Neighborhood:
* a Multifaceted Collaborative Filtering Model",
* available at [[http://public.research.att.com/~volinsky/netflix/kdd08koren.pdf]].
*
* The prediction rule is rui = u + bu + bi + qi*(pu + |N(u)|^^-0.5^^*sum(y)),
* see the details on page 6.
*
* @param edges edges for constructing the graph
*
* @param conf SVDPlusPlus parameters
*
* @return a graph with vertex attributes containing the trained model
*/
def run(edges: RDD[Edge[Double]], conf: Conf)
: (Graph[(Array[Double], Array[Double], Double, Double), Double], Double) =
{
// Generate default vertex attribute
def defaultF(rank: Int): (Array[Double], Array[Double], Double, Double) = {
// TODO: use a fixed random seed
val v1 = Array.fill(rank)(Random.nextDouble())
val v2 = Array.fill(rank)(Random.nextDouble())
(v1, v2, 0.0, 0.0)
}
// calculate global rating mean
edges.cache()
val (rs, rc) = edges.map(e => (e.attr, 1L)).reduce((a, b) => (a._1 + b._1, a._2 + b._2))
val u = rs / rc
// construct graph
var g = Graph.fromEdges(edges, defaultF(conf.rank)).cache()
materialize(g)
edges.unpersist()
// Calculate initial bias and norm
val t0 = g.aggregateMessages[(Long, Double)](
ctx => { ctx.sendToSrc((1L, ctx.attr)); ctx.sendToDst((1L, ctx.attr)) },
(g1, g2) => (g1._1 + g2._1, g1._2 + g2._2))
val gJoinT0 = g.outerJoinVertices(t0) {
(vid: VertexId, vd: (Array[Double], Array[Double], Double, Double),
msg: Option[(Long, Double)]) =>
(vd._1, vd._2, msg.get._2 / msg.get._1 - u, 1.0 / scala.math.sqrt(msg.get._1))
}.cache()
materialize(gJoinT0)
g.unpersist()
g = gJoinT0
def sendMsgTrainF(conf: Conf, u: Double)
(ctx: EdgeContext[
(Array[Double], Array[Double], Double, Double),
Double,
(Array[Double], Array[Double], Double)]) {
val (usr, itm) = (ctx.srcAttr, ctx.dstAttr)
val (p, q) = (usr._1, itm._1)
val rank = p.length
var pred = u + usr._3 + itm._3 + blas.ddot(rank, q, 1, usr._2, 1)
pred = math.max(pred, conf.minVal)
pred = math.min(pred, conf.maxVal)
val err = ctx.attr - pred
// updateP = (err * q - conf.gamma7 * p) * conf.gamma2
val updateP = q.clone()
blas.dscal(rank, err * conf.gamma2, updateP, 1)
blas.daxpy(rank, -conf.gamma7 * conf.gamma2, p, 1, updateP, 1)
// updateQ = (err * usr._2 - conf.gamma7 * q) * conf.gamma2
val updateQ = usr._2.clone()
blas.dscal(rank, err * conf.gamma2, updateQ, 1)
blas.daxpy(rank, -conf.gamma7 * conf.gamma2, q, 1, updateQ, 1)
// updateY = (err * usr._4 * q - conf.gamma7 * itm._2) * conf.gamma2
val updateY = q.clone()
blas.dscal(rank, err * usr._4 * conf.gamma2, updateY, 1)
blas.daxpy(rank, -conf.gamma7 * conf.gamma2, itm._2, 1, updateY, 1)
ctx.sendToSrc((updateP, updateY, (err - conf.gamma6 * usr._3) * conf.gamma1))
ctx.sendToDst((updateQ, updateY, (err - conf.gamma6 * itm._3) * conf.gamma1))
}
for (i <- 0 until conf.maxIters) {
// Phase 1, calculate pu + |N(u)|^(-0.5)*sum(y) for user nodes
g.cache()
val t1 = g.aggregateMessages[Array[Double]](
ctx => ctx.sendToSrc(ctx.dstAttr._2),
(g1, g2) => {
val out = g1.clone()
blas.daxpy(out.length, 1.0, g2, 1, out, 1)
out
})
val gJoinT1 = g.outerJoinVertices(t1) {
(vid: VertexId, vd: (Array[Double], Array[Double], Double, Double),
msg: Option[Array[Double]]) =>
if (msg.isDefined) {
val out = vd._1.clone()
blas.daxpy(out.length, vd._4, msg.get, 1, out, 1)
(vd._1, out, vd._3, vd._4)
} else {
vd
}
}.cache()
materialize(gJoinT1)
g.unpersist()
g = gJoinT1
// Phase 2, update p for user nodes and q, y for item nodes
g.cache()
val t2 = g.aggregateMessages(
sendMsgTrainF(conf, u),
(g1: (Array[Double], Array[Double], Double), g2: (Array[Double], Array[Double], Double)) =>
{
val out1 = g1._1.clone()
blas.daxpy(out1.length, 1.0, g2._1, 1, out1, 1)
val out2 = g2._2.clone()
blas.daxpy(out2.length, 1.0, g2._2, 1, out2, 1)
(out1, out2, g1._3 + g2._3)
})
val gJoinT2 = g.outerJoinVertices(t2) {
(vid: VertexId,
vd: (Array[Double], Array[Double], Double, Double),
msg: Option[(Array[Double], Array[Double], Double)]) => {
val out1 = vd._1.clone()
blas.daxpy(out1.length, 1.0, msg.get._1, 1, out1, 1)
val out2 = vd._2.clone()
blas.daxpy(out2.length, 1.0, msg.get._2, 1, out2, 1)
(out1, out2, vd._3 + msg.get._3, vd._4)
}
}.cache()
materialize(gJoinT2)
g.unpersist()
g = gJoinT2
}
// calculate error on training set
def sendMsgTestF(conf: Conf, u: Double)
(ctx: EdgeContext[(Array[Double], Array[Double], Double, Double), Double, Double]) {
val (usr, itm) = (ctx.srcAttr, ctx.dstAttr)
val (p, q) = (usr._1, itm._1)
var pred = u + usr._3 + itm._3 + blas.ddot(q.length, q, 1, usr._2, 1)
pred = math.max(pred, conf.minVal)
pred = math.min(pred, conf.maxVal)
val err = (ctx.attr - pred) * (ctx.attr - pred)
ctx.sendToDst(err)
}
g.cache()
val t3 = g.aggregateMessages[Double](sendMsgTestF(conf, u), _ + _)
val gJoinT3 = g.outerJoinVertices(t3) {
(vid: VertexId, vd: (Array[Double], Array[Double], Double, Double), msg: Option[Double]) =>
if (msg.isDefined) (vd._1, vd._2, vd._3, msg.get) else vd
}.cache()
materialize(gJoinT3)
g.unpersist()
g = gJoinT3
// Convert DoubleMatrix to Array[Double]:
val newVertices = g.vertices.mapValues(v => (v._1.toArray, v._2.toArray, v._3, v._4))
(Graph(newVertices, g.edges), u)
}
/**
* Forces materialization of a Graph by count()ing its RDDs.
*/
private def materialize(g: Graph[_, _]): Unit = {
g.vertices.count()
g.edges.count()
}
}
|
