They can't be random was such, because some users had around 8 ST or 8 GK lists from around 10-12 recommendations. The probability of having such outcomes is very low (near 0). If you consider the recommendations list types as 4 main categories, GK, DF, MF, ST, the probability of having a list of one category is 1/4. The probability of having the same list category for 5 times in a row is (1/4 x 1/4 x 1/4 x 1/4 x 1/4) = 0.00098. Now, just calculate how low the probabilities are for the outcomes that some users had. Based on observations, there are some randomness involved in the recommendations' feature, but it is unlikely that randomness are among the main factors for deciding the category of the list that will appear.


810Likes
LinkBack URL
About LinkBacks

