Прогнозирование с использованием идентификатора пользователя в Recommenderlab
У меня довольно простая проблема с рекомендациями: я хотел бы создать гибридный рекомендатель с recommenderlab в R, где я рекомендую уже понравившиеся фильмы в MovieLense набор данных вместе с популярными фильмами и некоторыми случайными фильмами.
library(recommenderlab)
data("MovieLense")
MovieLense100 <- MovieLense[rowCounts(MovieLense) > 100,]
train <- MovieLense100[1:100]
test <- MovieLense100[101:103]
## mix popular movies with a random recommendations for diversity and
## rerecommend some movies the user liked.
recom <- HybridRecommender(
Recommender(train, method = "POPULAR"),
Recommender(train, method = "RANDOM"),
Recommender(train, method = "RERECOMMEND"),
weights = c(.6, .1, .3)
)
Когда я обучил модель, я хотел бы "предсказать" фильмы для данного пользователя, используя userId. Документация Recommenderlab предполагает, что это возможно, если идентификатор пользователя находится в данных обучения. Тем не менее, когда я пытаюсь предсказать, используя userid в newdata Я получаю следующее сообщение об ошибке:
as(predict(object = recom, newdata = 1), "list")
#> Error in object@predict(object@model, newdata, n = n, data = data, type = type, : If newdata is a user id then data needes to be the training dataset.
Справедливо, попробую с вектором:
as(predict(recom, newdata = c(1,5)), "list")
#> Error in object@predict(object@model, newdata, n = n, data = data, type = type, : If newdata is a user id then data needes to be the training dataset.
Не могу понять, что означает это сообщение об ошибке. Кто-нибудь знает?
1 ответ
Решение
Вам просто нужно уточнить, что вы рекомендуете элементы пользователю или пользователям из учебного набора данных следующим образом:
> as(predict(recom, 1, data = train), "list")
$`1`
[1] "Hearts and Minds (1996)" "Boys, Les (1997)" "Great Day in Harlem, A (1994)"
[4] "Dangerous Beauty (1998)" "Tough and Deadly (1995)" "Two or Three Things I Know About Her (1966)"
[7] "Santa with Muscles (1996)" "Margaret's Museum (1995)" "Brassed Off (1996)"
[10] "Thousand Acres, A (1997)"
> as(predict(recom, 1:10, data = train), "list")
$`1`
[1] "Dangerous Beauty (1998)" "Great Day in Harlem, A (1994)" "Hearts and Minds (1996)"
[4] "Boys, Les (1997)" "Santa with Muscles (1996)" "Two or Three Things I Know About Her (1966)"
[7] "Tough and Deadly (1995)" "Saint of Fort Washington, The (1993)" "Whole Wide World, The (1996)"
[10] "Margaret's Museum (1995)"
$`5`
[1] "Boys, Les (1997)" "Two or Three Things I Know About Her (1966)" "Santa with Muscles (1996)"
[4] "Dangerous Beauty (1998)" "Great Day in Harlem, A (1994)" "Whole Wide World, The (1996)"
[7] "Hearts and Minds (1996)" "Brassed Off (1996)" "Saint of Fort Washington, The (1993)"
[10] "Celestial Clockwork (1994)"
$`6`
[1] "Dangerous Beauty (1998)" "Two or Three Things I Know About Her (1966)" "Santa with Muscles (1996)"
[4] "Tough and Deadly (1995)" "Brassed Off (1996)" "Crooklyn (1994)"
[7] "Great Day in Harlem, A (1994)" "Hearts and Minds (1996)" "Saint of Fort Washington, The (1993)"
[10] "Boys, Les (1997)"
$`7`
[1] "Two or Three Things I Know About Her (1966)" "Tough and Deadly (1995)" "Santa with Muscles (1996)"
[4] "Dangerous Beauty (1998)" "Hearts and Minds (1996)" "Great Day in Harlem, A (1994)"
[7] "Boys, Les (1997)" "Brassed Off (1996)" "Margaret's Museum (1995)"
[10] "Crooklyn (1994)"
$`10`
[1] "Dangerous Beauty (1998)" "Great Day in Harlem, A (1994)" "Hearts and Minds (1996)"
[4] "Tough and Deadly (1995)" "Two or Three Things I Know About Her (1966)" "Boys, Les (1997)"
[7] "Santa with Muscles (1996)" "Brassed Off (1996)" "Margaret's Museum (1995)"
[10] "Crooklyn (1994)"
$`11`
[1] "Two or Three Things I Know About Her (1966)" "Hearts and Minds (1996)" "Great Day in Harlem, A (1994)"
[4] "Tough and Deadly (1995)" "Brassed Off (1996)" "Dangerous Beauty (1998)"
[7] "Boys, Les (1997)" "Santa with Muscles (1996)" "Thousand Acres, A (1997)"
[10] "Horse Whisperer, The (1998)"
$`13`
[1] "Santa with Muscles (1996)" "Brassed Off (1996)" "Tough and Deadly (1995)" "Boys, Les (1997)"
[5] "Thousand Acres, A (1997)" "Margaret's Museum (1995)" "Total Eclipse (1995)" "Love! Valour! Compassion! (1997)"
[9] "Withnail and I (1987)" "Prefontaine (1997)"
$`15`
[1] "Two or Three Things I Know About Her (1966)" "Boys, Les (1997)" "Great Day in Harlem, A (1994)"
[4] "Tough and Deadly (1995)" "Dangerous Beauty (1998)" "Saint of Fort Washington, The (1993)"
[7] "Santa with Muscles (1996)" "Hearts and Minds (1996)" "Celestial Clockwork (1994)"
[10] "Whole Wide World, The (1996)"
$`16`
[1] "Great Day in Harlem, A (1994)" "Dangerous Beauty (1998)" "Hearts and Minds (1996)"
[4] "Two or Three Things I Know About Her (1966)" "Boys, Les (1997)" "Santa with Muscles (1996)"
[7] "Tough and Deadly (1995)" "Brassed Off (1996)" "Crooklyn (1994)"
[10] "Thousand Acres, A (1997)"
$`18`
[1] "Great Day in Harlem, A (1994)" "Two or Three Things I Know About Her (1966)" "Dangerous Beauty (1998)"
[4] "Hearts and Minds (1996)" "Santa with Muscles (1996)" "Tough and Deadly (1995)"
[7] "Boys, Les (1997)" "Whole Wide World, The (1996)" "Margaret's Museum (1995)"
[10] "Saint of Fort Washington, The (1993)"
В противном случае, предсказать функцию мы не знаем, на какого пользователя вы ссылаетесь.
Вы также можете прогнозировать данные испытаний следующим образом:
> as(predict(recom, test), "list")