Sistema_recomendaciones

Librerías

library(tidyverse)   

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.4     ✔ readr     2.1.5
## ✔ forcats   1.0.0     ✔ stringr   1.5.1
## ✔ ggplot2   3.5.0     ✔ tibble    3.2.1
## ✔ lubridate 1.9.3     ✔ tidyr     1.3.1
## ✔ purrr     1.0.2     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(recommenderlab)

## Loading required package: Matrix
## 
## Attaching package: 'Matrix'
## 
## The following objects are masked from 'package:tidyr':
## 
##     expand, pack, unpack
## 
## Loading required package: arules
## 
## Attaching package: 'arules'
## 
## The following object is masked from 'package:dplyr':
## 
##     recode
## 
## The following objects are masked from 'package:base':
## 
##     abbreviate, write
## 
## Loading required package: proxy
## 
## Attaching package: 'proxy'
## 
## The following object is masked from 'package:Matrix':
## 
##     as.matrix
## 
## The following objects are masked from 'package:stats':
## 
##     as.dist, dist
## 
## The following object is masked from 'package:base':
## 
##     as.matrix
## 
## Registered S3 methods overwritten by 'registry':
##   method               from 
##   print.registry_field proxy
##   print.registry_entry proxy

Matriz de evaluaciones de películas

movie_ratings_tarea <- data.frame(
  panda = c(5, 4, 5, 4), 
  enredados = c(4, 3, 4, NA),
  UP = c(4, NA, 4, 5),
  CARS = c(5, 5, 5, NA),
  Toystory  = c(3, 2, 3, 4),
  shrek = c(5, 3, 4, 5),
  monja= c(2, 4, 2, NA)
)
rownames(movie_ratings_tarea) <- c('Luis','David',  'Emilio', 'Yo' )
movie_ratings_tarea

##        panda enredados UP CARS Toystory shrek monja
## Luis       5         4  4    5        3     5     2
## David      4         3 NA    5        2     3     4
## Emilio     5         4  4    5        3     4     2
## Yo         4        NA  5   NA        4     5    NA

cosine_similarity <- function(vec_1, vec_2) {
  vec_len <- length(vec_1)
  
  # NA values are replaced with 0
  vec_1[is.na(vec_1)] <- 0
  vec_2[is.na(vec_2)] <- 0
  
  # Computing the denominator
  vec_1_denom <- sqrt(sum(vec_1^2))
  vec_2_denom <- sqrt(sum(vec_2^2))
  denominator <- vec_1_denom * vec_2_denom
  
  # Computing the numerator
  tib = tibble(vec_1 = vec_1, vec_2 = vec_2)
  tib <- tib %>% mutate(products = vec_1 * vec_2)
  numerator <- sum(tib$products)
  
  # Return the cosine similarity
  return (numerator / denominator)
}

# Obtenemos los vectores de cada persona
You <- as.numeric(as.vector(movie_ratings_tarea['Yo',]))
Luis <- as.numeric(as.vector(movie_ratings_tarea['Luis',]))
David <- as.numeric(as.vector(movie_ratings_tarea['David',]))
Emilio <- as.numeric(as.vector(movie_ratings_tarea['Emilio',]))


# Obtenemos "distancia" usando similitud por conseno
similarities_tarea <- data.frame(
  cosine_similarity = c(cosine_similarity(You, David), cosine_similarity(You, Luis), cosine_similarity(You, Emilio))
)
rownames(similarities_tarea) <- c('David', 'Luis', 'Emilio')
similarities_tarea

##        cosine_similarity
## David          0.4845562
## Luis           0.7762349
## Emilio         0.7546822

# Creamos una función para obtener un promedio ponderado en base a los amigos
movie_rating_weighted_average <- function(movie, friends) {
  denominator <- 0
  numerator <- 0
  for (friend in friends) {
    friend_similarity <- similarities_tarea[friend,][1]
    friend_rating <- movie_ratings_tarea[friend, movie][1]
    
    # Tomaremos el promedio ponderado en cuenta solo para películas con calificación
    if (is.na(friend_rating)) next
    
    denominator <- denominator + friend_similarity
    numerator <- numerator + (friend_similarity * friend_rating)
  }
  
  return (numerator / denominator)
}

friend_names <- c('Luis', 'David', 'Emilio')
new_movies <- c('enredados', 'CARS', 'monja')
new_movie_predicted_ratings <- tibble()
for (n in new_movies) {
  predicted_rating <- movie_rating_weighted_average(n, friend_names)
  prediction_tibble <- tibble(movie = n, predicted_rating = predicted_rating)
  new_movie_predicted_ratings <- bind_rows(new_movie_predicted_ratings, prediction_tibble)
}
new_movie_predicted_ratings

## # A tibble: 3 × 2
##   movie     predicted_rating
##   <chr>                <dbl>
## 1 enredados             3.76
## 2 CARS                  5   
## 3 monja                 2.48