user_based_recom.RdUse the user item matrix to search for similar users in order to get the anime they graded the best and that we have not watched
user_based_recom( userid = 999999999, user_item_matrix, ratings_data, n_recommendation = 5, threshold = 1, nearest_neighbors = 10 )
| userid | user id, set to 999999999 so that there is no conflict with other Id's |
|---|---|
| user_item_matrix | matrix created by the function `user_item_matrix` |
| ratings_data | table of ratings similar to the one we kept |
| n_recommendation | number of recommendations wanted |
| threshold | threshold level of number of user that gave a score to the anime |
| nearest_neighbors | number of neighbors taken into account for the computation |
Return a table composed of `n_recommendation` that the user have not seen yet
Marie Bellier, Massimo Finini, Meri Likoska, Vania Rodrigues Telo Ramos, Xavier Renger
#Simple example to showcase what is happening in the application #import dataset from the package anime_with_ratings <- tibble::tibble(ProjectG5::anime_with_ratings) #selection of the user example names_selected <- c("Death Note", "Naruto") score_entered <- c(10,7) temp_tibble <- tibble::tibble(Name = names_selected, rating = score_entered) anime_selected <- dplyr::left_join(anime, temp_tibble, by = c("Name" = "Name")) anime_selected <- dplyr::filter(anime_selected, Name %in% names_selected) anime_selected <- dplyr::mutate(anime_selected, user_id = 999999999) #creation of the matrix with what the user selected item_matrix <- user_item_matrix(anime_with_ratings, adding_row = TRUE, row_data = anime_selected) user_based_recom(999999999, item_matrix, anime_with_ratings, 5, 1, 10) #> # A tibble: 5 x 6 #> # Groups: item_id, Name, Episodes [5] #> item_id Name Episodes Duration count rating #> <chr> <chr> <chr> <int> <int> <dbl> #> 1 ID2001 Gurren Lagann 27 24 3 10 #> 2 ID6 Trigun 26 24 3 10 #> 3 ID7311 The Disappearance of Haruhi Suzumiya 1 162 3 10 #> 4 ID2025 Darker than Black 25 24 2 10 #> 5 ID1 Cowboy Bebop 26 24 3 9.67