Data understanding, preparation, and recommendations

The one library that we will need for this exercise is recommenderlab. The package was developed by the Southern Methodist University's Lyle Engineering Lab, and they have an excellent website with supporting documentation at https://lyle.smu.edu/IDA/recommenderlab/:

> library(recommenderlab)

> data(Jester5k)

> Jester5k
5000 x 100 rating matrix of class 'realRatingMatrix' with
362106 ratings.

The rating matrix contains 362106 total ratings. It is quite easy to get a list of a user's ratings. Let's look at user number 10. The following output is abbreviated for the first five jokes:

> as(Jester5k[10,], "list")
$u12843
   j1    j2    j3    j4    j5 ...
-1.99 -6.89  2.09 -4.42 -4.90 ...

You can also look at the mean rating for a user (user 10) and/or the mean rating for a specific joke (joke 1), as follows:

> rowMeans(Jester5k[10,])
u12843 
  -1.6

> colMeans(Jester5k[,1])
  j1 
0.92

One method to get a better understanding...