recommend
li_volleyball
2016年3月20日
library(recommenderlab)
library(ggplot2)
#
data(MovieLense)
dim(MovieLense)
## [1] 943 1664
MovieLense
## 943 x 1664 rating matrix of class 'realRatingMatrix' with 99392 ratings.
image(sample(MovieLense,500),main="Raw ratings")
qplot(getRatings(MovieLense),binwidth=1,main="histogram of normalized ratings",xlab = "Ratings")
summary(getRatings(MovieLense))
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.00 3.00 4.00 3.53 4.00 5.00
#normalized ratings
qplot(getRatings(normalize(MovieLense,method="Z-score")),main="hist of normalized ratings",xlab="rating")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
summary(getRatings(normalize(MovieLense,method="Z-score")))
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -4.8520 -0.6466 0.1084 0.0000 0.7506 4.1280
qplot(rowCounts(MovieLense), binwidth=10, main="Movies rated on Average", xlab="# of users", ylab="# of movies rated")
qplot(colMeans(MovieLense), binwidth=0.1, main="Mean ratings of Movies", xlab="Rating", ylab="# of movies")
recommenderRegistry$get_entries(dataType="realRatingMatrix")
## $IBCF_realRatingMatrix
## Recommender method: IBCF
## Description: Recommender based on item-based collaborative filtering (real data).
## Parameters:
## k method normalize normalize_sim_matrix alpha na_as_zero minRating
## 1 30 Cosine center FALSE 0.5 FALSE NA
##
## $PCA_realRatingMatrix
## Recommender method: PCA
## Description: Recommender based on PCA approximation (real data).
## Parameters:
## categories method normalize normalize_sim_matrix alpha na_as_zero
## 1 20 Cosine center FALSE 0.5 FALSE
## minRating
## 1 NA
##
## $POPULAR_realRatingMatrix
## Recommender method: POPULAR
## Description: Recommender based on item popularity (real data).
## Parameters: None
##
## $RANDOM_realRatingMatrix
## Recommender method: RANDOM
## Description: Produce random recommendations (real ratings).
## Parameters: None
##
## $SVD_realRatingMatrix
## Recommender method: SVD
## Description: Recommender based on EM-based SVD approximation from package bcv (real data).
## Parameters:
## approxRank maxiter normalize minRating
## 1 NA 100 center NA
##
## $UBCF_realRatingMatrix
## Recommender method: UBCF
## Description: Recommender based on user-based collaborative filtering (real data).
## Parameters:
## method nn sample normalize minRating
## 1 cosine 25 FALSE center NA
scheme <- evaluationScheme(MovieLense, method="split", train=0.9, k=1, given=10, goodRating=4)
scheme
## Evaluation scheme with 10 items given
## Method: 'split' with 1 run(s).
## Training set proportion: 0.900
## Good ratings: >=4.000000
## Data set: 943 x 1664 rating matrix of class 'realRatingMatrix' with 99392 ratings.
algorithms <- list(
"random items" = list(name="RANDOM", param=list(normalize = "Z-score")),
"popular items" = list(name="POPULAR", param=list(normalize = "Z-score")),
"user-based CF" = list(name="UBCF", param=list(normalize = "Z-score", method="Cosine", nn=50, minRating=3)),
"item-based CF" = list(name="IBCF", param=list(normalize = "Z-score", method="Cosine"))
)
# run algorithms, predict next n movies
results <- evaluate(scheme, algorithms, n=c(1, 3, 5, 10, 15, 20))
## RANDOM run fold/sample [model time/prediction time]
## 1 [0.02sec/1.13sec]
## POPULAR run fold/sample [model time/prediction time]
## 1 [0.14sec/0.2sec]
## UBCF run fold/sample [model time/prediction time]
## 1 [0.11sec/52.33sec]
## IBCF run fold/sample [model time/prediction time]
## 1 [348.01sec/0.66sec]
plot(results, annotate = 1:4, legend="topleft")
# See precision / recall
plot(results, "prec/rec", annotate=3)
summary(results)
## Length Class Mode
## random items 1 evaluationResults S4
## popular items 1 evaluationResults S4
## user-based CF 1 evaluationResults S4
## item-based CF 1 evaluationResults S4
print(results)
## List of evaluation results for 4 recommenders:
## Evaluation results for 1 folds/samples using method 'RANDOM'.
## Evaluation results for 1 folds/samples using method 'POPULAR'.
## Evaluation results for 1 folds/samples using method 'UBCF'.
## Evaluation results for 1 folds/samples using method 'IBCF'.
library(plyr)
result1<-ldply(avg(results))
head(result1)
## .id TP FP FN TN precision
## 1 random items 0.00000000 1.000000 47.75789 1605.242 0.00000000
## 2 random items 0.05263158 2.947368 47.70526 1603.295 0.01754386
## 3 random items 0.09473684 4.905263 47.66316 1601.337 0.01894737
## 4 random items 0.23157895 9.768421 47.52632 1596.474 0.02315789
## 5 random items 0.32631579 14.673684 47.43158 1591.568 0.02175439
## 6 random items 0.48421053 19.515789 47.27368 1586.726 0.02421053
## recall TPR FPR
## 1 0.000000000 0.000000000 0.0006231881
## 2 0.000420633 0.000420633 0.0018345901
## 3 0.001343461 0.001343461 0.0030535159
## 4 0.002965187 0.002965187 0.0060813035
## 5 0.004276282 0.004276282 0.0091353054
## 6 0.007966717 0.007966717 0.0121507535
result1[,1]<-paste(result1[,1],c(1, 3, 5, 10, 15, 20))
temp_result1<-result1[,c(1,6,7)]
f<-function(p,r){
return(2*p*r)/(p+r)
}
result1_f<-cbind(result1,f=f(temp_result1[,2],temp_result1[,3]))
head(result1_f)
## .id TP FP FN TN precision
## 1 random items 1 0.00000000 1.000000 47.75789 1605.242 0.00000000
## 2 random items 3 0.05263158 2.947368 47.70526 1603.295 0.01754386
## 3 random items 5 0.09473684 4.905263 47.66316 1601.337 0.01894737
## 4 random items 10 0.23157895 9.768421 47.52632 1596.474 0.02315789
## 5 random items 15 0.32631579 14.673684 47.43158 1591.568 0.02175439
## 6 random items 20 0.48421053 19.515789 47.27368 1586.726 0.02421053
## recall TPR FPR f
## 1 0.000000000 0.000000000 0.0006231881 0.000000e+00
## 2 0.000420633 0.000420633 0.0018345901 1.475905e-05
## 3 0.001343461 0.001343461 0.0030535159 5.091011e-05
## 4 0.002965187 0.002965187 0.0060813035 1.373350e-04
## 5 0.004276282 0.004276282 0.0091353054 1.860558e-04
## 6 0.007966717 0.007966717 0.0121507535 3.857568e-04
head(result1_f[order(-result1_f$f),])
## .id TP FP FN TN precision
## 18 user-based CF 20 6.094737 12.273684 41.66316 1593.968 0.3381538
## 17 user-based CF 15 4.978947 8.915789 42.77895 1597.326 0.3629917
## 16 user-based CF 10 3.684211 5.684211 44.07368 1600.558 0.3948758
## 12 popular items 20 5.368421 14.631579 42.38947 1591.611 0.2684211
## 11 popular items 15 4.421053 10.578947 43.33684 1595.663 0.2947368
## 15 user-based CF 5 2.157895 2.610526 45.60000 1603.632 0.4532609
## recall TPR FPR f
## 18 0.16566075 0.16566075 0.007591384 0.11203762
## 17 0.13829264 0.13829264 0.005510458 0.10039817
## 16 0.10324963 0.10324963 0.003508480 0.08154156
## 12 0.12821289 0.12821289 0.009037106 0.06883008
## 11 0.11281484 0.11281484 0.006530564 0.06650138
## 15 0.06796729 0.06796729 0.001609646 0.06161383
#带入模型
moive_re<-Recommender(MovieLense,method="UBCF")
moives_pr<-predict(moive_re,MovieLense,n=20)
print(as(moives_pr,"list"))
## [[1]]
## [1] "Glory (1989)"
## [2] "Schindler's List (1993)"
## [3] "Casablanca (1942)"
## [4] "Close Shave, A (1995)"
## [5] "Dr. Strangelove or: How I Learned to Stop Worrying and Love the Bomb (1963)"
## [6] "Leaving Las Vegas (1995)"
## [7] "One Flew Over the Cuckoo's Nest (1975)"
## [8] "Rear Window (1954)"
## [9] "Heathers (1989)"
## [10] "L.A. Confidential (1997)"
## [11] "City of Lost Children, The (1995)"
## [12] "Butch Cassidy and the Sundance Kid (1969)"
## [13] "Titanic (1997)"
## [14] "Lawrence of Arabia (1962)"
## [15] "Shine (1996)"
## [16] "Stand by Me (1986)"
## [17] "Gandhi (1982)"
## [18] "To Kill a Mockingbird (1962)"
## [19] "In the Name of the Father (1993)"
## [20] "Harold and Maude (1971)"
##
## [[2]]
## [1] "Boot, Das (1981)"
## [2] "Dead Man Walking (1995)"
## [3] "Lone Star (1996)"
## [4] "Return of the Jedi (1983)"
## [5] "Celluloid Closet, The (1995)"
## [6] "Casablanca (1942)"
## [7] "Citizen Kane (1941)"
## [8] "Godfather: Part II, The (1974)"
## [9] "2001: A Space Odyssey (1968)"
## [10] "When We Were Kings (1996)"
## [11] "Diva (1981)"
## [12] "Close Shave, A (1995)"
## [13] "Tango Lesson, The (1997)"
## [14] "Beautiful Thing (1996)"
## [15] "Empire Strikes Back, The (1980)"
## [16] "Mrs. Dalloway (1997)"
## [17] "Butch Cassidy and the Sundance Kid (1969)"
## [18] "My Fair Lady (1964)"
## [19] "Bonnie and Clyde (1967)"
## [20] "Annie Hall (1977)"
##
## [[3]]
## [1] "Mrs. Brown (Her Majesty, Mrs. Brown) (1997)"
## [2] "Star Wars (1977)"
## [3] "Pulp Fiction (1994)"
## [4] "English Patient, The (1996)"
## [5] "Full Monty, The (1997)"
## [6] "Lone Star (1996)"
## [7] "Titanic (1997)"
## [8] "Sweet Hereafter, The (1997)"
## [9] "In the Company of Men (1997)"
## [10] "Willy Wonka and the Chocolate Factory (1971)"
## [11] "In & Out (1997)"
## [12] "Vertigo (1958)"
## [13] "As Good As It Gets (1997)"
## [14] "Apt Pupil (1998)"
## [15] "Dazed and Confused (1993)"
## [16] "Ice Storm, The (1997)"
## [17] "This Is Spinal Tap (1984)"
## [18] "Trainspotting (1996)"
## [19] "Heat (1995)"
## [20] "Fargo (1996)"
##
## [[4]]
## [1] "Titanic (1997)" "English Patient, The (1996)"
## [3] "L.A. Confidential (1997)" "Game, The (1997)"
## [5] "Good Will Hunting (1997)" "Kiss the Girls (1997)"
## [7] "Full Monty, The (1997)" "Usual Suspects, The (1995)"
## [9] "Rosewood (1997)" "Boogie Nights (1997)"
## [11] "Raise the Red * (1991)" "Pulp Fiction (1994)"
## [13] "Toy Story (1995)" "Love Jones (1997)"
## [15] "Eve's Bayou (1997)" "Edge, The (1997)"
## [17] "Sting, The (1973)" "Some Like It Hot (1959)"
## [19] "Strictly Ballroom (1992)" "Soul Food (1997)"
##
## [[5]]
## [1] "Terminator 2: Judgment Day (1991)"
## [2] "Terminator, The (1984)"
## [3] "Usual Suspects, The (1995)"
## [4] "Contact (1997)"
## [5] "Braveheart (1995)"
## [6] "Casablanca (1942)"
## [7] "Twelve Monkeys (1995)"
## [8] "Godfather, The (1972)"
## [9] "Shawshank Redemption, The (1994)"
## [10] "Raising Arizona (1987)"
## [11] "Amadeus (1984)"
## [12] "Nikita (La Femme Nikita) (1990)"
## [13] "Reservoir Dogs (1992)"
## [14] "Citizen Kane (1941)"
## [15] "Dr. Strangelove or: How I Learned to Stop Worrying and Love the Bomb (1963)"
## [16] "Schindler's List (1993)"
## [17] "Titanic (1997)"
## [18] "Leaving Las Vegas (1995)"
## [19] "North by Northwest (1959)"
## [20] "Army of Darkness (1993)"
#S4 class 是一种标准的R语言面向对象实现方式，s4对象有明确的类定义，参数定义，参数检查，继承关系，实例化等面向对象系统的特征。