Ratings For Baselines?

For an accurate prediction, we need to relate the available information of a movie with the preferences of a user. In case there is not much information available, or in the worst case, none, a mean rating can be used as a baseline to indicate what the crowd thinks about the movie. However, as mentioned earlier, the mean rating might have a strong bias depending on the movie and many other latent factors. Thus, a value of 6.8 can make perfect sense for one user, because the agreement with the crowd is usually very high, but others might be very disappointed because they expected something else.

We tried to follow mean ratings of some recent blockbuster movies and compared them with the (expert) reviews and the resulting ‘stars’ the movie would get. The analysis was not very deep, but there is a strong indicator that such movies get an extra boost just for the popularity which might be partly explained by well known actors or top notch special effects.

In a nutshell, the mean ratings tend to be always >6.0 even if there are (serious) flaws people agree about, because the crowd gave good ratings because they enjoyed the popcorn cinema which is perfectly okay. The point is that something as diverse as a movie can be hardly condensed in a single value that is useful for a broad range of people because the expectations of users are also very diverse. Maybe it would be useful to ask users “Was this rating useful” to use the feedback to estimate the utility for further mean ratings…


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s